Skip to main content

Full text of "Neoplastic diseases, a text-book on tumors"

See other formats


The original of this book is in 
the Cornell University Library. 

There are no known copyright restrictions in 
the United States on the use of the text. 




JAMES EWING, A. M., M. D.. Sc. D. 

Professor of Pathology at Cornell University Medical College 
New York City 






Copyright, 1919, by W. B, Saunders Company 



It is the object' of this work to present within reasonable space and in 
accessible form the main features of the origin, structure, and natural history 
of tumors. 

Up to a very recent time it has been the prevailing impression that 
tumors faU into a limited number of grand classes in which the forms occur- 
ring in the several organs are so nearly related as to be virtually identical. 
Hence the practical physician or surgeon has been content to regard all 
fibromas, sarcomas, or cancers, as equivalent conditions without regard to 
the organ involved, and on this theory to treat the members of each class 
aUke. Upon this theory also it was legitimate to conceive of a universal 
causative agent of malignant tumors and thus to subordinate many very 
obvious differences which clinical experience has established in the origin 
and behavior of different related tumors. 

1 believe that this point of view has greatly retarded the progress of the 
knowledge of tumors, and it has been the writer's effort to combat such a 
conception, so far as present knowledge permits. He has endeavored to 
analyze the numerous etiologic factors which meet in such diverse fashions 
in the inception of tumors, to emphasize the general dependence of clinical 
course upon histologic structure, to trace the histogenesis to the last de- 
gree, impressing its essential importance when known, and to enumerate 
and contrast the more striking chnical features which are often highly 
characteristic of different tumors. 

No one would think of confusing lobar pneumonia with pneumonic 
plague, although both are examples of acute exudative pneumonitis, but it is 
quite the rule to identify for statistical studies several equally different forms 
of mammary cancer. The former diseases are related only as forms of 
inflammation, the latter only as types of neoplasia. From this point of 
view it may safely be said that there are more distinct clinical and pathologic 
entities within the groups of neoplasms than exist outside of them. 

While a great volume of information regarding the clinical phenomena of 
the main forms of tumors is available in special works on medicine, surgery, 
and the specialties, the task of unraveling their separate varieties, tracing 
their mode of origin and growth, and establishing the nature of the less com- 
mon forms, falls to the lot of the pathologist. For the final classification 
of tumors must depend chiefly on histogenesis and structure. Present-day 
oncology is chiefly concerned with these topics and the space devoted 
to them can safely be reduced only when our knowledge is much further 

In spite of several laborious years spent in the task the writer acknowl- 
edges disappointment with the results attained in many departments, but 
can only claim that the effort to present tumors as specific diseases is in the 
right direction. He first undertook to write a book on the general principles 
of oncology, but soon found that the significant facts about tumors are not 
of general application, but are best revealed in the study of special tumor 
groups or even of special cases. 

In the compilation of the work the writer has endeavored to consult with 



due respect all the standard authorities, and as far as possible the original 
contributions in the literature. The rather extensive bibliographic lists 
seem necessary for the guidance of the reader who desires complete informa- 
tion and to whom the work is chiefly addressed. The recent rapid increase 
in original contributions from the United States has made it impossible to 
do full justice to American literature. 

It is a pleasure to acknowledge the valuable assistance received from 
many friends and colleagues here and abroad. The author acknowledges 
especially his indebtedness to Messrs. W. B. Saunders Company for under- 
taking the publication of a highly specialized work. Through the pains- 
taking assistance of Mr. William Dunn it has been possible to rely almost 
entirely upon photographs for illustrations. 

While confessing a deep interest in the theoretic problems which render 
oncology the most complex and fascinating field in pathology, the chief 
object and hope of the author have been that by rendering more accessible 
to English readers the knowledge of tumors he may contribute something 
toward the reduction of the mortality from cancer. 

James Ewing. 
January, 1919. 





Historical - ... ... . . 17 


DEriNiTiON. Classification. General Pathology . 25 


Malignancy and Its ErPECTS on the Organism ... 62 

Metastasis . 76 


Chemistry of Tumors; Serology . . , 89 


Theories of the Nature of Cancer. 94 


The Special Etiology of Tumors; Trauma. . 109 


The Parasitic Theory .... 114 


Experimental Cancer Research. . 127 


Fibroma . ..... ... 149 






. i66 




Chondroma . 


. 180 

Osteoma. . 



Myoma . 



Angioma. . . 







Clinical Types oe Sarcoma . 

Sarcomas of Bone and Bone-marrow .... 



Lymphoma and Lymphosarcoma . 


Tumors ojf the Brain. ... . . 





• 378 




Tumors OP Nerve Trunks . . 411 

Tumors of Spinal Cord and Membranes. . 419 

General Pathology OF Epithelial Tumors. ... . 429 

Epithelial AND Other Tumors OF THE Breast . . . ...... 467 

Cancer of Uterus, Vulva, Vagina -523 

Chorioma (Chorionepithelioma) . 546 

Cysts and Epithelial Tumors of the Ovary . 562 

The Ovarian Teratoma ... . . 593 

Carcinoma of Stomach . 605 

Carcinoma of Intestine . 640 

Epithelial Hyperplasia and Tumors of Liver . . . 654 

Tumors of Pancreas . . 681 

Maxillary Tumors of Dental Origin . . . . 686 

Epithelial Tumors of the Salivary Glands . . . . 704 

Tumors of Kidney 717 




Tumors ojf the Adrenal • 74^ 

Tumors of Prostate 7S7 

Tumors of Testis 768 

Tumors of Lung 7^5 

Epidermoid Carcinoma .... . . 797 


Epidermoid Carcinoma {Continued) . . . 835 

Melanoma . . . . 849 

Tumors of the Thyroid . . . . 869 

The Thymus and Its Tumors .... 889 

Tumors of the Hypophysis ... 901 

The Pineal Gland and Its Tumors 925 

Teratology . . . 931 

Bibliography 963 

Index 1007 





The Ancients knew cancer well. They treated it by excision and by a 
variety of escharotics, including the Egyptian arsenical ointment. Cancer is 
mentioned in the Papyrus Ebers (B.C. 1500) and in the oldest remnants of 
the literature of India and Persia. 

Hippocrates (B.C. 460-375) received from earlier days a considerable 
body of descriptive facts regarding cancer of the skin, breast, uterus and 
internal organs, and he first employed the terms /cap/ctijos for all indolent 
ulcers, and KapKivcciia for progressive malignant tumors. The humoral 
pathology then disseminated conceptions of the origin of cancer. Deficiency 
or excess of blood, mucus or bile, formed the basis of all disease. Herodotus 
mentions that Democedes (B.C. 520) cured Atossa, the daughter of Darius 
Hystaspis, of breast cancer, and Hippocrates burnt out a carcinoma of the 
neck, the earliest record of diathermia. 

Celsus (B.C. 30-A.D. 38) distinguished several gross varieties of cancer, 
and he excised breast cancer, advising against removal of the pectoralis 
major. Treatment by charcoal was employed by Cato, and a variety of 
crude internal remedies are mentioned by Pliny. 

Galen (A.D.'1i3i-203), the founder of experimental physiology and 
pathology, failed to make any significant advance in the conception of 
cancer, but the presentation of the humoral doctrine of atra bilis in his 
writings formed a Scripture which dominated medical thought for more than 
a thousand years. Swelhngs were secundam naturam (gravid uterus), supra 
naturam (callus formations), or prater naturam (true tumors). The pneuma, 
■composed of solid parts and four fluids, blood, mucus, yellow and black 
bile, ruled the processes of the body. Cancer developed from the concen- 
tration of black bile. 

Suppression of menses and hemorrhoids, preventing the discharge of black 
bile, were chiefly responsible for cancer, which appeared where the bile gravi- 
tated in face, Ups, breast, etc. No modern writer has been able to reflect 
the thought of that period or to explain the firm entrenchment of the crude 
humoral theories. Since capable logic could not have been lacking, one 
must suppose that religious and esthetic tendencies in the race unfitted the 
human mind for natural thought regarding the structure and functions of 
human body. For internal cancer, of which little was known, a diet, chiefly 
vegetable was recommended. Walnuts were specifically forbidden. 
2 17 


Diagnosis rested chiefly on the cause of the disease, while treatment 
by excision, ligation of vessels, and cautery was comparatively successful. 
Leonides of Alexandria (A.D. i8o) broke away from Hippocrates' conserva- 
tism, dissected out breast cancer extensively, cutting through healthy tissue 
with knife and cautery and approached closely to the modern technics of 
this operation. 

In the Byzantine period (475^1500) considerable progress was made in 
the description of various tumors. Paulus of Aegina (625-690) separated 
chronic metritis from uterine cancer. In Arabia, Avicenna (980-1037) intro- 
duced the internal use of arsenic, and Avenzoar (1070-1162) employed the 
esophageal sound and the nutrient enema. During the i2th-i5th centuries 
the ban of the Church was extended from dissection of the body to the prac- 
tice of surgery which could be studied only in private and practised only by 
nomads, while the monks gave themselves over to translations of Galen's 
writings and to speculations in alchemy, astrology and magic. With the 
opening of universities at Paris (mo), Salerno (1150), Montpellier (1150), 
and Prag (1348) , the Church dictated what books should be used. 

The Renaissance (i 500-1 700) bringing the discovery of the printing press 
and the circulation of the blood (Harvey, 1628) greatly facihtated the spread 
of knowledge and aided the more accurate diagnosis and better treatment of 
cancer, but threw httle light on etiology. Andreas Vesal (1514-1564) began 
the attack on many of the concepts of Galen, identifying deep seated with 
ulcerating cancer. Fabricius (1537-1619) separated many inflammatory 
swelUngs from cancer, warned rigidly against incomplete removal, extirpated 
the uterus, and experimented with internal remedies. 

Marcus Aurelius Severinus (1580-1656) described myxosarcoma, sepa- 
rated cancer from benign tumors of the breast, and extirpated the axillary 

In Prag, Sennert (1572-1637) and in Lisbon, Lusitanus (1642) first claimed 
that cancer was contagious, a doctrine which prevailed extensively and 
with little dissent. Paracelsus (1413-1541) stands out as the first successful 
opponent of Galen's theory of atra hilis as the cause of cancer. He claimed 
that the disease was due to mineral salts in the blood. He seemed to think 
that cancer developed where various animal salts became concentrated and 
sought an outlet. At this period it is interesting to note that the decline 
of Galen's authority and distrust of even his crude theories of etiology led 
to complete demoralization in the treatment of cancer, encouraged great 
abuse of arsenic and other internal and external remedies, permitted the 
faith cure career of Queen EHzabeth (1602), and developed many fantastic 
theories regarding the nature of cancer. 

The Lymph Theory. — In the 17th century Galen's doctrine was completely 
demolished by the discovery of the circiilation by Harvey (1628), of the 
lymph-vessels by Glens (1652), and of the red blood-cells by Malpighi (1661). 

Malpighi used the microscope which had been in existence since 1592. 
The black bile was nowhere to be found but everywhere was blood and lymph. 
Lymph coagulated and foamed on boiling, hence cancer was composed of 
lymph varying in density, alkalescence, or acidity, and in maUgnant tumors 
fermenting and degenerating. Louis (1723-92) distinguished gelatinous 
lymph (goiter) and albuminous lymph (scirrhus). LeDran (1685-1770) 
studied cancer by many autopsies, emphasized the local nature of cancer of 
the skin, and the internal origin of breast cancer. He conceived that if a 
drop of cancer lymph passed the adjacent nodes it contaminated the entire 

Astruc (1684-1766) separated cysts from true tumors, showed that 


scirrhus and soft cancer were of the same nature, pointed out important 
differences in prognosis of different types of carcinoma, and by incineration 
proved that cancer and muscle tissue contained the same salts. Yet he 
likened the growth of cancer from lymph to the heating and swelhng of 
gypsum in water. 

Morgagni (1682-1772) estabHshed the importance of the pathological 
anatomy of cancer, describing many internal tumors studied at autopsy 
and separating gumma, struma; exostosis, and lipoma from cancer. He 
distrusted, without replacing, the lymph theory. 

A notable event in the history of cancer research was the essay of Peyrilhe 
(i73S""i8o4) submitted to the Academy of Lyons in answer to the question, 
Qu' est ce que le cancer? This was the first systematic investigation of the 
whole subject and dealt with the cancer toxin, the nature of the disease, the 
manner of growth and the treatment. 

He spoke of the local origin, the production of a specific virus from 
degeneration of the tumor and the development of cachexia from this source. 

He endeavored to demonstrate the virus by injecting tumor emulsion 
beneath the skin of a dog but an abscess resulted and his servant drowned 
the animal. He treated ulceration effectively with the newly discovered 
carbonic acid. His imagination was satisfied by the Cartesian lymph theory 
of the times. 

In Germany the 17th century was occupied by the exploitation of many 
personal views of the origin of cancer and other diseases. 

Chemical conceptions held sway, and cancer was attributed usually to 
excess of acid, to be treated by alkali (Helmont, Ettmiiller). 

Stahl (1660-1734), supposed that stasis and thickening of the blood 
were the essential factors. Hoffmann held that life and health depended 
on normal movements of the tissues, cancer and other diseases resulting 
from atony, stasis, and abnormal fermentation of blood and lymph. He 
constructed his anodyne as a panacea to restore tissue tone. 

In England, John Hunter's (1728-93) conception of the lymph theory 
represented a distinct advance toward the cellular pathology. He held that 
tumors grew from the coagulable lymph constantly thrown out of the blood, 
that tumor tissue like normal tissue was nourished by the organism, and 
developed according to the same biological laws. In 1802, The Society for 
investigating the Nature and Cause of Cancer was formed in London, and 
formulated the problems of the disease as they stand to-day. The Society 
dissolved in 1806. 

The pathological anatomy of malignant tumors chiefly interested the 
English observers, and Hey (i 736-1819) described in detail the structure 
of certain vascular tumors which he found to be composed of organized 
blood-fluid and a large proportion of lymph, and for which he employed the 
term "fungus hematodes." Wardrop, 1809, described in great detail miscella- 
neous tumors of this gross type, attempting to separate them from cancer. 
Many of his cases occurred in children, some involved the eye, and one arose 
from a wart. 

Abernethy, 1804, attempted to define the old term sarcoma by applying 
it to various soft tumors, and he called Hey's fungus hematodes, medullary 
sarcoma. Maunoir, 1820, showed the beneficent result of anatomical study, 
by announcing that each tumor is the result of a morbid change of the fluid 
or tissue from which it arises, and retains the original characters of this 
tissue. He employed many crude chemical and physical tests to support 
this claim. 

The close attention then being given to clinical data led Pott (1775) to 


recognize and describe chimney-sweeps' cancer and to point out its etiology. 
The English contributions of this period were completed by Home, who, 
using the microscope, described and depicted rounded bodies which he 
regarded as lymph corpuscles. 

In France, Bichat's studies of tissue structure opened a new era with the 
appearance of his Anatomic generate in 1801. He first distinguished the 
stroma from the parenchyma of tumors, deriving the former from the origi- 
nating tissue and the latter by proliferation from the stroma. Without reli- 
ance upon the microscope he regarded the stroma as cellular and conceived of 
its growth by proliferation. 

Following Bichat, Laennec made a systematic study of the gross features 
of cancer in thin shces, and introduced the term "encephaloid".for the soft 
parts. Dupuytren attempted to prove the infectiousness of the disease by 
ingestion and intravenous inoculation of cancer tissue. Bayle and Cayol 
argued in favor of the constitutional nature of the disease and fully pointed 
out the difference between chronic mastitis and cancer. 

Broussais' doctrine that cancer was the sequel of recurrent inflammation 
had made some influence on the thought of this period. Based on chemical 
observation, his claim that cancer never arises in normal tissues but only 
after inflammatory alterations, found many adherents and is of interest in 
connection with later theories. 

Lobstein used the term "plastic lymph" and formulated the view that the 
tumor-forming lymph was not under control of the biological laws of the 
organism. Recamier studied especially the infiltration of cancer, observed 
the destruction of veins and applied the term "metastases" to nodules in 
the brain in cases of mammary cancer. He recognized the importance of 
supernumerary organs and nevi as sources of cancer. Andral reflected the 
confusion of ideas then prevailing by offering his theory that products of 
secretion became organized into cancer. Cruveilhier regarded cancer as a 
malignant degeneration which like inflammation affected all the organs. Its 
chief pathognomonic feature was the cancer juice exuded on pressure. Andral 
thought tumor masses floating in the veins were derived from the fibrin, but 
Velpeau, in a case of intravenous tumor, searched the blood in vain for can- 
cerous elements. Such questions as the identity of encephaloid or soft cancer 
with brain tissue were hardly settled. Tubercle was commonly confused 
with cancer and gummata imperfectly distinguished from it. Cancers were 
classified upon crude gross characters as in Jager's Handworterbuch, Leip- 
sic, 1837, where they were divided in four types, hard, soft, pigmented, and 

Thus during the sway of the lymph theory, English and French students 
while adding important contributions to the descriptive history and gross 
pathology of raahgnant tumors, failed to pass beyond the limits of the pre- 
vailing theoretical conceptions of the time. It was the great period of indi- 
vidual surmises which in some instances, notably with Hunter, Home, Lob- 
stein and Recamier, approached closely toward and prepared the way for 
modern conceptions. 

In Germany, Richter, Walther and others engaged in the current discus- 
sions without adding essentially new data. 

It was the general conception that the elements of cancer were fluid and 
traveled in the veins. 

Great significance was attached to the observation of tumor masses in the 
vessels. A special variety of "blood-cancer" supposed to come from irritat- 
ing elements was described by Langstaff, 1817, and Carswefl, 1834. 

Histological Period.^ — With the construction of the achromatic micro- 


scope in Paris, 1824, a new era opened in cancer research. Wolf finds that 
the first fruitful studies of the structure of vegetable and animal tissues by 
this instrument were made by Raspail in 1826, who showed that the growth 
of tissues resulted from the multiplication of cells. He clearly stated the 
doctrine of the cell, finding that tissues were composed of microscopic vesicles. 
The structure and growth of fat tissue he described in detail. Collard, 1828, 
also described rather clearly the stages in the embryonal development*' of 
tissues assuming however, that the cells originated from plastic lymph. 
Schwann, 1838, established this doctrine of cellular structure as a universal 
principle and discovered the nucleus and nucleolus of the cell. 

In the same year J. MiiUer published his classical study of malignant 
tumors. He found them all to be composed of groups of cells, each contain- 
ing nuclei and nucleoli. The various current types of cancer he found to be 
distinguished only by different proportions and groupings of cell masses and 
stroma. Certain elongated or racquet-shaped cells " geschwanzten Kbrper- 
chen" he regarded as !on the way to fiber formation, but not as specific can- 
cer elements which he was unable anywhere to detect. Hence his diagnosis 
of cancer rested on clinical signs and the anatomical grouping of the cells. 
He held the interesting view that cancer developed not from normal tissue 
but from germ cells which as a "seminium morbi" lay scattered between the 
tissue elements. 

In regard to the origin of the cells it was held that most of them de- 
veloped from the plastic lymph or blastema by a process of budding. Others 
resulted from division of nuclei and cell body, or by the transformation of 
intracellular blastema lying in spaces within the cell body (Virchow's Brut- 
raume). The originating tumor-cells were not derivatives of the normal 
tissue cells but came from the seminium morbi, hence there was urgent de- 
mand and search for specific characters in the tumor-cells. Lebert especially 
described such specific cells and designated as pseudocancer all tumors, as 
rodent ulcer, which failed to contain them. Following this principle, Hann- 
over carefully described the group of tumors arising from stratified squamous 
epithelium and separated them from cancer under the term "epithehoma." 

Hannover believed that cancer-cells circulated in the blood and produced 
metastases as pus-cells produced pyemia. 

The doctrine of the specific cancer-cell now became the chief topic of dis- 
cussion and was supported by many writers. It was opposed by Bruch 
who added many details to Miiller's work, by Virchow who found the tailed 
corpuscles in the normal bladder epithelium, and by Velpeau who with 
Hannover, established the microscopical diagnosis of tumors. Endogenous 
cell formation then took the place denied the "tailed corpuscle" as the 
pathognomonic sign of cancer, Bruch, Virchow, Remak and many others 
accepting the importance of this supposed type of cell growth. The signi- 
ficance of tumor stroma also attracted new attention, especially from Rokit- 
ansky who explained many of the gross features from variations in this ele- 
ment. He also studied the secondary degenerations of cancer tissue, 
including inflammation, necrosis, and saponification, and considered the 
possibility of a spontaneous cure by these processes. 

In spite of their very careful histological studies of tumor tissue'in the 
fresh condition all the authoritative writers of this period including Virchow 
were led to beheve in the origin of cancer from a fluid blastema. Cancer 
was defined as an organized exudate from the blood with overnutrition and 
overgrowth. Much controversy arose concerning the various^ types of 
blastemas supposed to exist and the changes in the blood from which they 
were all necessarily derived. Vogel thought there were as many blastemas 


as tissues, different tumors arising in different tissues according to the type 
or analogy of the tissue involved (law of analogous blastemas). These 
views led directly to the conception of cancer as a constitutional dyscrasia 
(Rokitansky), a belief that gained wide acceptance and seemed to be sup- 
ported by the anemia of the disease. The help of the chemists was sought 
to separate from the blood the different blastemas and Fiihrer was able_ to 
distinguish albuminous, chondrinous and glutinous varieties, each of which 
gave a suitable color reaction with nitric acid, and which were offered as 
sources of the corresponding tumors. The idea that exuded elements of the 
blood could become organized into cancer was vigorously opposed by Cruveil- 
hier who urged that extravasated blood never became organized but had 
lost all claim to vitahty. Cancer could therefore develop only in the vessels. 
This reasoning led Langenbeck to study cancerous masses in the veins which 
he found to be composed of fibrin, pus-cells, and cancer-cells. He therefore 
drew the important conclusion that cancer-cells possess a remarkable ca- 
pacity for independent existence and that they were carried through the veins 
producing secondary tumors or metastases. After Peyrilhe, Alibert, 1806, 
and Dupuytren, 1817, produced only suppuration by intravenous injections 
of tumor emulsion, but Langenbeck, Follin and Velpeau claimed to have 
observed tumor nodules in the lungs of dogs receiving intravenous injections. 
These results, although faihng to receive confirmation seemed then to prove 
not only the origin of metastases but the contagiousness of cancer. Yet 
Bruch at once interpreted them as transportation of cancer-cells and not as 
infection. He held the lymph-vessels to be the chief channels of transport 
and with Meckel described backward transport in lymphatics". 

While still burdened by the blastema theory of origin the study of cancer 
had succeeded up to i860 in rather accurately describing and classifying 
the main classes of tumors chiefly according to microscopical structure. The 
description of the benign tumors, the existence of various types of carcinoma, 
the malignancy of epithehoma and the separate position of sarcomas, were 
generally accepted facts. Correct conceptions of the histogenesis however 
were impossible until Virchow founded the cellular pathology upon the doc- 
trine of Omnis cellula e cellula. There had been several opponents of the 
idea that cells could be formed from exuded lymph, notably Cruveilhier, 
while Remak at the same time with Virchow claimed that cells grew ex- 
clusively from other cells by endogenous reproduction; but Virchow appUed 
the new principle rigidly to all departments and especially to the origin and 
growth of tumors. Coincidently with this memorable service he fell into 
two grave errors. He failed to correctly interpret the deceptive evidence of 
endogenous cell formation and he was led to believe that cancer-cells origin- 
ated from connective-tissue cells. The latter error he never fully rehnquished 
and possibly on this account his monumental work "Die krankhaften Gesch- 
wiilste" was never extended over the field of cancer. 

Remak immediately attacked the connective tissue theory which had 
become very popular but which conflicted with the significant principle of the 
immutabihty of the three separate germ layers which he with His had already 
established. Remak insisted that epithelial cells could arise only from 
epithelium. The appearance of such cells in abnormal places he referred 
to the misplacement of embryonal cell groups. 

Virchow's authority however prevented the general acceptance of 
Remak's views for many years, until the appearance of Thiersch's famous 
monograph on epitheUal cancer. 

Although many authors opposed the theory of the connective origin of 
cancer, and Meckel in 1857 had traced the origin of a buccal tumor directly 


from the lining epithelium, it remained for Thiersch to present convincing 
evidence of the invariable derivation of epithelioma from lining epithelium. 
By improvements in hardening, cutting and staining the tissues in serial 
sections he traced the growth of several epithehomas from the Malpighian 
layer or glandular structures of the skin and demonstrated the principle, 
omnis cellula e cellula ejusdem generis. He also placed great importance upon 
previous changes in the epithelial connective tissue as precipitating the down- 
ward growth of epithehum. 

By a process of involution associated with diminished nutrition, functional 
capacity and mechanical resistance, the histogenetic equilibrium between 
stroma and epithelium became disturbed and overgrowth of epithehum 
resulted. This was the first competent physiological conception of the patho- 
genesis of cancer and with it Thiersch introduced the modern era of our 
knowledge of the nature of the disease. Waldeyer extended Thiersch's 
observations to the internal organs and traced the origin of cancer of stomach, 
liver and kidney to the epithehal cells of these organs. He first described 
the isolation of these cells by indurated connective tissue and held that 
tumors developed from these isolated cells. The formation of secondary 
tumors he demonstrated to be the result of continuous growth through blood- 
and lymph-vessels as well as of cell emboli. 

The acceptance of the epithelial theory was briefly delayed by the studies 
of Recklinghausen and Koster, who brought forward evidence to show that 
many cancers arose from the endothelium of lymph spaces. The outcome 
of this controversy was the establishment of the group of endothehoma. 
That true cancer has no connection with lymphatic endothelium was shown 
by Carmalt who in Waldeyer's laboratory shook thin sections of fresh cancer 
in silver solution, thereby removing the cancer-cells from the lymph spaces 
and leaving the silvered endothelium intact. Yet Virchow's influence still 
predominated and led to the assumption that cancer developed from almost 
any mesoblastic cell, as from sarcolemma (Popper), perimysium (R. Volk- 
mann), cartilage (Weber), vessel- wall (Gussenbauer) and leucocytes (Classen, 

Through the untiring industry of himself and his pupils Waldeyer slowly 
forced the acceptance of his views of the origin of cancer. Billroth slowly 
abandoned much of his early allegiance to the connective tissue doctrine. 
It was of great assistance when the regeneration of epithehum over wounds 
was shown to advance from the edges of intact skin. By the minute study 
of a large number of cancers of many varieties, Waldeyer eventually 
established the principle of the exclusive epithelial origin of cancer and of 
its growth from its own resources, and demohshed the former belief that 
normal tissues could become transformed into cancer. 

The classification of tumors upon the histogenetic basis was accomplished 
with the data furnished by the studies of this period. The theory of a 
cancerous dyscrasia necessarily lost ground with the demonstration that 
the disease has a local origin and that secondary tumors arise from trans- 
ported cells. Likewise it became evident that the parasitic theory which 
had enjoyed a general acceptation must be set aside as an inadequate 
explanation of the new facts of tumor origin and growth. 

Further than this the cellular pathology seemed incapable of carrying 
progress in the etiological explanation of cancer. Irritation, trauma and 
infection seemed to be connected with the origin of the disease but the obscure 
nature of the relation of«any of these factors was fully recognized. 

The last decades of the nineteenth century were thus occupied with the 
detailed study of the morphology of tumors, the separation of varieties of the 


disease, the elucidation of histogenesis, and the writing of the natural history 
of mahgnant diseases. The twentieth century opens as the experimental era 
with the systematic study of tumors throughout the animal kingdom, and it 
seems hkely to become noteworthy as the period of specific etiological 
investigation which promises to widely separate many neoplastic diseases 
formerly held to be closely related. It may thereby prove to be the era of 
successful therapeutics and prophylaxis. 


Definition. — A tumor is an autonomous newgrowth of tissue. 

Among most authors there is virtual agreement regarding the essential 
parts of the definition of tumors, v. Rindfleisch says briefly: "A tumor is a 
locahzed degenerative excess of growth." Ziegler defines a tumor as a new- 
growth of tissue which apparently originates and grows spontaneously, 
possesses an atypical structure, does not subserve the uses of the organism, 
arid reaches no definite termination of its growth. According to Birch- 
Hirschfeld, tumors are progressive newgrowths of tissue arising sponta- 
neously from cells of its own, and separated from normal tissue both in mor- 
phology and in function. Ribbert states that tumors are newgrowths of 
tissue, self-centered, largely or wholly uncontrolled by the organism which 
supplies only their nutrition, with a structure never entirely normal, and 
reaching no definite termination of growth. 

Lubarsch includes among tumors all spontaneously arising tissue growths 
which, while typical in form, differ in histology from the originating tissue, 
and in spite of organic connection with the body pursue an autonomous 
course rarely to the benefit of the whole body. Borst emphasizes in the 
tumor process the unknown etiology, the local overgrowth, the peculiar 
autonomy, the purposeless and endless course, and the atypical morphology 
and biology of the tumor product. Adami accepts White's descriptive 
definition: "A tumor proper is a mass of cells, tissues, or organs, resembling"! 
those normally present but arranged atypically. It grows at the expense ' 
of the organism without at the same time subserving any useful function." _j 

I beheve with Prudden that beyond the autonomy of tumor growth it 
is difficult to add any element to our definition which may apply- to all true 
blastomas. The more descriptive definitions are doubtless more readily 
applied but with increasing knowledge the unknown etiology, progressive 
growth, and even the atypical morphology, may become less significant, while 
the physiological conception of autonomy grows more substantial. Certain 
obvious features of tumor growth emphasize the importance of their autonomy. 

The nature of tumor growth is different from that of the normal tissues 
and from inflammatory hyperplasia, and is something new and foreign to the 
organism in which it occurs. 

In most tumors growth is progressive, and no natural termination is real- 
ized, the tumor growing most rapidly at the death of the patient. Yet the 
progressive quaUty is not invariable as some tumors spontaneously regress; 
others, as scirrhus cancer, long remain practically quiescent, and others de- 
generate or necrose or become extruded, while fibrosis overtakes a few. 

The growth of tumors is new to the organism in that it is usnaHy expansive, 
emanating from an isolated group of cells and pushing contiguous tissue be- 
fore it. Such tumors are more or less encapsulated, appearing as isolated 
or parasitic organoid structures. In some tumors the growth is infiltrative, 
single cells or cell groups pushing their way through and destroying adjacent 
tissues. In many tumors growth is extensive, beginning in a small focus and 
gradually precipitating the contiguous normal cells into the tumor process. 


Tumor-cells acquire new and greatly increased powers of growth often 
exceeding those possessed by any parallel normal tissue. 

Ribbert claims that tumor-cells do not reveal any abnormal capacity 
for growth since these are determined in the ovum, but merely exhibit their 
innate capacities because the restraints to growth have been lost. It is 
true that the organized regenerative properties of tissue cells are determined in 
the fertilized ovum, but the power of multipHcation without organization is 
subject to great variations from the environment, and observation and ex- 
periment show that tumor-cells have acquired in this sense abnormal capacity 
for proHferation. The growth of tumor-cells is peculiar in that it proceeds 
under abnormal conditions when the cells have been separated from their 
natural connections, and even when they are transferred to other animals of 
the same species. In transplanted tumors the cells often grow with increased 
activity, when normal cells rapidly necrose or undergo slow regression, or 
adjust themselves to the organization. Thus in many particulars the growth 
of tumors differs from the organfzed regeneration of normal tissues. 

These physiological peculiarities will later be considered in detail. 


There are three obvious and useful plans for the classification of tumors: 
histological, regional, and etiological. 

1. A thorough etiological classification of tumors cannot with our pres- 
ent knowledge be accompHshed with sufficient detail for most practical 
purposes. Yet many useful terms, which have more or less etiological 
significance, are employed to designate tumors. Most tumors are post- 
natal in origin; many are congenital, as the multiple chondromas of the spinal 
column, sarcoma of the kidney, and angioma of the skin. 

There are some notable tumors which show a striking hereditary character, 
as glioma of the retina. A large class of tumors of many varieties are appar- 
ently of traumatic origin. Partly etiological is the grouping of tumors into 
adult and embryonal, arising from adult or embryonal cells and reproducing 
adult or embryonal cells and structures. 

The list of known specific causes of tumors is small but enlarging, in- 
cluding the cancers of pipe-smokers, chimney-sweepers, and workers in 
paraffin, a;-ray cancers, Kangri cancer and that of chewers of betelnut among 
Africans. Some tumors probably result from exaggerated response to func- 
tional stimuh, as the thyroid tumors of fish; others from prolonged mechan- 
ical, chemical or thermal irritation; some are closely related to tuberculous, 
syphilitic and other infectious processes. As knowledge advances it seems 
probable that a much wider extension of etiological classification along these 
and other lines will be accomplished. 

2. The regional classification of tumors is of great practical utility but 
as a rule it only deals with superficial characters of neoplasms and usually re- 
solves itself into a grouping of distinct histological varieties occurring in 
different organs. Yet there are exceptions to the rule. Hypernephroma is 
a regional term referring to tumors chiefly of the head of the kidney but in- 
cluding several histological and structural types. Multiple myeloma varies 
in structure but is practically the sole primary tumor of bone-marrow. Ter- 
atoma of the testis and adamantinoma of the maxillae are almost the only 
tumors arising in these regions and both cover such a wide variety of struc- 
tures that it is convenient to employ regional terms to designate such tumors. 
Moreover the marked peculiarities in the origin and course of similar his- 
tological tumors in different organs often render it advisable to emphasize 


the organ involved rather than the structure of the tumors. The mor- 
phologist speaks of cancer of mamma, larynx, uterus, etc., as one and the same 
disease, while the clinician properly conceives of mammary cancer, laryn- 
geal cancer, uterine cancer, etc., as quite distinct clinical entities. There is 
much practical value in emphasizing the regional point of view, and as knowl- 
edge advances it may well prove that the strictly histological grouping of 
many tumors is based on structural resemblances which are less significant 
than are now supposed. 

3. The histological structure of tumors offers the simplest means of 

A tumor receives a name and position according to the tissue which it 
most resembles. Yet this plan meets difficulties at once and applies readily 
only to simple and benign tumors. Many malignant tumors do not resemble 
any nof mal tissue and terms drawn from gross pathology, sarcoma and carci- 
noma, have been drafted to designate the microscopical structure of the two 
main classes of tumors. 

Moreover to name a tumor according to its microscopical structure often 
conveys very little information about it. 

Tumor nomenclature aims to be more specific and there is a constant 
tendency to replace histological with histogenetic terms wherever possible. 
The main classes of tumors are therefore named from their histology, while 
the specific forms are designated, if possible, by their histogenesis, as neuro- 
fibroma, adenoma sebaceum, etc. 

Where structure varies, histogenesis dominates the nomenclature, as 
with adamantinoma, hypernephroma, etc. 

Histological Classification 

GROUP I.— TYPE: connective TISSUE 

(ff) Fibroma, composed of connective tissue. 

(b) Chondroma, composed of cartilage. 

(c) Chordoma, composed of tissue of chorda dorsalis. 

(d) Osteoma, composed of bone. 

(e) Myxoma, composed of mucous tissue. 
(/) Lipoma, composed of fat tissue. 

(g) Angioma, composed of blood-vessels. 
(h) Lymphoma, composed of lymphatic tissue. 

(i) Sarcoma, a cellular tumor composed of anaplastic tissue of any of the 
above types. 


(a) Leiomyoma, composed of smooth muscle tissue. 

(b) Rhabdomyoma, composed of striated muscle. 


(a) Neuroma, composed of nerve fibers. 

(b) Neuroma ganglionare, of nerve fibers and ganglion cells. 

(c) Glioma, composed of glia tissue. 

(d) Neuro-epithelioma, composed of neuro-epithelium. 


(a) Papilloma, a tumor of pavement epithehum, with supporting tissues in 
normal arrangement. 


(b) Adenoma, a benign tumor of glandular epithelium, with supporting 
tissue, in normal arrangement. 

(c) Epithelioma, or epidermoid carcinoma, epithelium in atypical 

- id) Carcinoma, a tumor of glandular epithelium in atypical arrangement. 


(a) Simple mixed tumors, composed of more than one type of neoplastic 
tissue — named according to composition: as chondro-epithelioma, adeno- 

(&) Teratoma, composed of tissues and organs of one, two, or three 
germinal layers, as monodermal, bidermal, or tridermal types. 

(c) Embryoma, composed of tissues from three germinal layers in more 
or less- orderly imitation of a fetus. 

The variable structure of many simple tumors requires complex terms 
for the accurate designation of structure. A fibroma with many neoplastic 
blood-vessels is called an angiofibroma, while a tumor chiefly composed of 
blood-vessels with much new connective tissue is called a fibro-angioma. 

A sarcomatous quality is indicated as fibro-, chondro-, or osteo-sarcoma. 
So fibro-adenoma and adenofibroma indicate predominance of one or another 
structure. Histioid tumors are composed of a simple tissue, organoid 
growths reproduce the structure of organs. Homologous tumors resemble 
the tissue in which they arise, as adenoma mammae; heterologous tumors 
present a structure foreign to the tissue in which they grow, as chondroma 

A strictly embryologic classification of tumors presents many advantages over others 
but for good reasons has never gained general adoption. 

The early segmenting embryo first consists of a mass (morula, blastula) of large un- 
differentiated cells (blastomeres). Soon these cells form two distinct layers, epiblast and 
hypoblast, both of which have the characters of lining membranes, the pavement quality 
being more marked in the former. Next the hypoblast gives rise to a group of cells forming 
the mesoblast and lying intermediate between the other two layers. The mesoblast cells 
lose their pavement qualities and appear in an undifferentiated cell mass, but, later, certain 
of its cells grow out between the epiblast and hypoblast in modified pavement form and 
enclose the body cavity (coelorn) . Thus the mesoblast comes to possess a lining portion, 
mesothelium, and a pulp portion, mesenchyme. 

The ccelom with its mesothelial lining eventually becomes the pleura, pericardium 
and peritoneum. From a portion of the mesothelium is also developed the parent tissue 
of the striated muscles, and later modified mesoblastic cells form the lining cells of the 
blood and lymphatic vessels. Thus the lining cells of the great serous cavities, and the 
vessels, and the striated muscles, are derived from mesoblastic cells which have at one time 
lost their pavement quality while merged in the mesoblast. 

At the same period with these changes, the epiblast proliferates in the region of the 
primitive groove and forces inward a mass of cells, which become isolated and modified, 
going to form the central nervous system. Coincidently, a portion of hypoblast retaining 
some of its pavement qualities arises to form the notochord. 

Adami points out that these early embryological processes lead to the development of 
two main classes of tissue which permanently retain their early charaaters, viz. : lining mem- 
brane tissues and pulp tissues, and which he proposes to call lepidic (Xcttis, membrane) 
and hylic (iiXjj, crude matter). Lepidic tissues are composed of groups of specific cells 
which are not penetrated by blood-vessels and possess no intercellular stroma. The epi- 
blast and hypoblast are primary, the mesothelium and endothelium are secondary lepidic 
tissues. Hylic or pulp tissues of which the mesenchyme is the type, are composed of cells 
separated by a homogeneous fibrillar stroma which may or may not be penetrated by 

Upon this basis Adami has divided the blastomas into two main groups: Lepidoma 
and Ilyloma. 

I. Lepidomata 

A. Primary, (i) Epilepidomata, typical (papilloma and adenoma), and at}rpical 
(carcinoma), all from epiblast. (2) Hypolepidomata, typical (adenoma) and atypical, 
(carcinoma) all from hypoblast. 


B. Secondary. (3) Mesolepidoraata, typical (adenoma of kidney, etc.) and atypical 
(carcinoma of kidney, etc.), all from mesothelium. (4) Endothelial lepidomata. 

II. Hylomata, or Pulp Tumors 

1. Epihylomata, typical (neuroma, glioma) and atypical (gliosarcoma) . 

2. HypoKylomata, chordoma. 

3. Mesohylomata, typical (fibroma, lipoma, etc.) and atypical (sarcoma) from mesenchy- 
mal tissues. Mesothelial hylomata or rhabdomyomata. 

There is no doubt that in the above scheme Adami has accomplished a successful 
classification and legitimate nomenclature for tissues and tumors, based upon accredited 
data. In several directions this classification is of value in the interpretation of many 
peculiarities of tumor growth. Thus the essential distinctions between lepidic or pavement 
tissues and hylic or pulp tissues established in the early embryo have a strong iiifluence 
upon tumors derived from these tissues. The degrees and types of the pavement quality 
observed in embryonal epiblast, hypoblast, and mesothelium reappear in much of the 
history of tumors derived from such tissues, especially the mesenchymal relations of en- 
dothelium. To a less extent, I think, the reversionary tendencies of tumor-ceUs show the 
influence of embryological relations. 

Yet there appear to me to be two main objections to the adoption of this most successful 
embryological classification and nomenclature: 

First, the behavior of tumor-cells is very much more influenced by the acquired char- 
acters of the cells of origin than by their embryological derivation. 

Second, the neoplastic process does not consist in retracing the steps of embryological 
development but probably in a progressive loss of original cell potencies, so that the cells 
become a prey to all manner of new external conditions which control their nutrition, 
greatly influence their form and render them significant chiefly as pathological parasitic 
ceUs and tissues. 

Embryology has been of great value in clearing up many of the disturbances of develop- 
ment which lead to the growth of heterologous tumors and teratomas, it aids in the inter- 
pretation of many phenomena of tumor growth, especially of teratomas, and will likely be 
of further aid in this field, but with the main element in the neoplastic process, anaplasia, 
it has nothing to do. 

Oncology is not a department of embryology but a separate chapter in the biology of 
the cell. Hence strictly embryological classification and nomenclature have never gained 
great favor, because they do not serve the largest needs of the student of tumors. The 
generally accepted plan of classification and terminology which is based on histology, 
modified as much as possible by histogenesis, is a natural product which has become 
very firmly established and probably deserves to prevail against the varying prominence 
of embryology, chemistry, and etiology. 


The external appearances of tumors present extremely wide variations 
far exceeding those of normal structures. Yet these gross features are as a 
rule distinctive, so that the recognition of a tumor may usually be made 
by inspection and palpation of the tissue. 

The terms sarcoma and carcinoma refer to the exuberant fleshy character 
of connective tissue growths and the rough similarity of cancer to the crab. 

The local circumscription of many tumors is a sufficient gross sign of many 
blastomas, but fails to separate mihary or nodular granulomas from many 
forms of multiple tumors. There are some cases of diffuse carcinoma, as 
of the breast, or carcinosis, as of the meninges, in which no visible tumor is 
produced, and only a minute attention to color, consistence, and distribution 
saves the gross diagnostician from error. 

The color of tumor tissue is often distinctive. 

The normal color of tissues is almost invariably altered, becoming opaque, 
lighter and often with a tinge of yellow or white, where groups of cancer-cells 
are growing. Virchow used to emphasize three zones in a cancer nodule, a 
central light or yellowish area of degeneration or cicatrization, a middle, 
firm, opaque, cellular zone, and a peripheral reddish area of congestive re- 
action. The congestion which marks the advance of many cancers in the 
skin may simulate a bacterial process. 


Blood content and hemorrhage chiefly determine the color of most tumors 
as of other tissues. Most cancers are comparatively bloodless while sarcomas 
are usually more vascular and often contain extravasated blood. Angiomas 
and angiosarcomas can usually be recognized in the gross from the congeries 
of blood-vessels composing them. Some sarcomas of bone are traversed by 
wide blood sinuses surrounded by little tumor tissue, and deserve the name of 
bone aneurism. Multiple hemorrhagic sarcoma of the skin (Kaposi) is 
marked~by frequent hemorrhages and pigmentation. Certain types of 
carcinoma, as chorioma, and teratoid cancer of the testis usually produce 
notably hemorrhagic tumors. 

The fungus hematodes of older writers is a cauliflower-like growth rich 
in blood sinuses. Hemorrhagic infarction may overtake many tumors 
causing them to assume a very dark red or rusty color. Secondary pig- 
mentation of large portions of tumors may result from altered extravasated 
blood. Carcinoma of the breast in dogs is nearly always pigmented from 
this source. 

Melanoma of the skin and choroid is a brown or black tumor containing 
autochthonous pigment, but in rapidly growing metastases it may pro- 
gressively lose its pigment. The greenish yellow color of chloroma is a dis- 
tinctive feature possibly caused by Hpochromes. 

A pearl gray color is seen in chondroma, myxoma, and some scirrhus 
cancers. Hypernephroma, except in its highly embryonal types and in 
rapidly growing metastases, has an ochre yellow tint. 

The consistence yields significant indications of the nature of many 
tumors. Osteoma and chondroma reveal themselves at once by stony hard- 
ness, but hard fibroma may simulate chondroma. 

The peculiar density of many carcinomatous nodules in the breast and 
of the bases and borders of carcinomatous ulcers is their chief diagnostic 
featiue. There are few organs in which the density of carcinoma is equaled 
by any other process, but gumma of the testis is usually firmer than the em- 
bryonal cancer of this organ. 

Many tumors are softened by edema, hemorrhage, necrosis, mucous 
degeneration and cyst formation, and some become very hard from fibrosis 
or calcification. In each organ carcinoma is usually firmer than adenoma 
and sarcoma, which have less fibrous tissue and more vessels. Rapidly 
growing and embryonal cancers fail to show the density of older t5rpes, and 
from this and other characters are difficult to distinguish in the gross from 
sarcoma. The soft elasticity of lipoma and especially of myxoma and 
colloid tumors is somewhat characteristic. Glioma is usually softer and more 
vascular than adjacent brain tissue. 

Cystic tumors are tense but yielding or fluctuating. Angiomas may 
be compressible. Lymphosarcoma is rather less firm than Hodgkin's 
granuloma, leukemia, or slow tuberculosis of lymph-nodes. 

Mucous carcinoma of peritoneum produces jelly-like infiltration or huge 
intraperitoneal masses, and endothelioma of pleura may yield peculiar dry 
crumbly material. The "sand tumors" of the dura (endothehoma) and 
ovary (adenoma), are easily recognized on palpation, and the waxy 
plates of cholesteatoma and sebaceous contents of dermoids are equally 

The texture and markings revealed by inspection of the surface and 
section of tumors often disclose their true nature. 

The firm translucent surface of a fibroma, the glistening homogeneous 
luster of a chondroma, the intertwined fascicles of muscle fibers in a myoma, 
and the branching twigs of a papillary tumor, are readily recognizable 


naked-eye features. Likewise the jelly-like material in myxoma and col- 
loid cancer, the whorls of many endotheliomas of the dura, the vascular 
twigs of angiosarcoma, and the lobules of fat tissue in lipoma furnish a 
ready basis for gross diagnosis. 

In dealing with cellular tumors there are many criteria which lead one to a 
correct conclusion. 

Very cellular tumors without much stroma are invariably soft and fragile 
or pultaceous (encephaloid) . Soft but coherent tumor masses usually show 
abundant stroma. Cellular tumors whether sarcoma or carcinoma are 
usually lighter colored and more opaque than the surrounding normal tis- 
sues. As a rule, sarcomas are more uniform in texture than carcinomas. 
The former may produce wide areas of new tissue unbroken by bands of 
connective or areas of necrosis although frequently interrupted by hemor- 
rhages. The latter exhibit the tendency of carcinoma to excite the growth 
of connective tissue and to undergo focal fatty degeneration which produce 
points and masses of yellowish or whitish material. The necrosis of sar- 
coma is usually more massive. 

Vertical section through an acanthoma shows compact columnar masses 
or whorls of cells, of pearly white lusler which even when infiltrating are well 
banked against the surrounding tissue. Early ulceration is also character- 
istic of this common tumor, but it rarely produces bulky masses, either 
primary or secondary. 

There are peculiar forms of carcinoma, as in the pylorus and rectum, 
marked by mucous degeneration and fibrosis, which are sometimes more 
safely identified in the gross than by the microscope. They produce annular 
constrictions, or wide infiltrations of the submucosa, but most of the cells 
may be lost by mucous degeneration. 

By naked eye inspection one can frequently recognize the main charac- 
ters of tumor structure many of which, from too exclusive reliance on the 
microscope, have come to possess chiefly a histological significance. 

The growth of a tumor may be mainly central so that it displaces the 
surrounding tissue by lateral pressure, as in myoma uteri. 

Central growth is well shown in some angiosarcomas in which actively 
growing blood-vessels sheathed with tumor-cells compose the central por- 
tions, while in the periphery the growth is sohd and grades into a firm fibrous 
capsule. Or the growth may be peripheral in which case the central portions 
are apt to be fibrosed or degenerated and depressed. 

Infiltrative growth is revealed by lack of capsule or other demarcation 
and by opaque cords of tumor-cells stretching out into the lymph spaces 
or vessels. Extensive intravascular growth characterizes hypernephroma, 
adenoma of liver, and chorioma. Polypoid, papillary and dendritic, 
and cystic architectures reveal themselves best to the naked eye, while 
more minute structures designated as fascicular, plexiform, and alveolar 
can usually be detected by careful inspection. 

Soft cellular tumors are often called medullary, or encephaloid, from 
resemblance to brain tissue, to which the counterpart is the hard scirrhus. 

Thrombosis of blood-vessels, edema, fatty, hyaline, mucoid and colloid 
degeneration, calcification and ossification, are all encountered in extreme 
degrees in the gross anatomy of tumors. 

Cysts of varying origin, form, and content are notable gross features of 
many tumors. 

Retention cysts filled with retained secretion develop in many adenomas 
especially of the ovary, some of which are the largest tumors ever observed. 
Originating in a closed group of alveoli whose secreting function is active 



dilatation of alveoli is accompanied by complex systems of papillary in- 
growths which constantly add their quota of secretion. Eventually in- 
flammatory exudate and hemorrhage are added with rapid increase in volume. 
The contents of such cysts are therefore serous fluid with much nucleo-albumin 
and mucus, or with pus and disintegrated chocolate colored blood. The 
walls are hned with varying quantities of papillary tumor growth. 

Liquefaction cysts with smooth walls form in fibromyomas, sarcomas, etc., 
from local edema, necrosis, and autolysis of solid elements. Their contents 
are thin and serous or mingled with granular or fatty detritus, blood or pus; 
or the fluids are removed by absorption. Extravasation cysts form after 
hemorrhages and exhibit all stages of alteration and absorption of blood. 
Dilated lymph- or blood-vessels contribute to the structure of some cystic 

Fig. I.- 

-Fibrosing epidermoid carcinoma. Spindle-shaped tumor-cells, 
infiltrative growth. 


sarcomas of bone and other tissues. Cystosarcoma phylloides is an old term 
sometimes applied to tumors showing a branching system of more or less 
parallel cysts recalling the veins of a leaf. 

It is often difficult to distinguish between true cystic blastomas and 
simple cysts with inflammatory overgrowth of hning cells. 

In most simple retention cysts there is a tendency for the epithelium 
to grow up into papillary projections, as in chronic mastitis, and with several 
forms of encysted animal parasites. Such inflammatory overgrowth may 
reach a considerable grade but usually lacks the distinctive morphology 
of a tumor, and regresses when the cyst is evacuated. Yet some of them may 
go on to produce genuine neoplasms. 



Ulceration overtakes most tumors of the skin and mucous membranes. 
The inadequacy and fragiHty of blood-vessels, degeneration and autolysis 
of tumor-cells, trauma, mechanical pressure and tension, contact with irri- 
tants causing inflammation, and bacterial infection, are the chief factors in 
its origin. 

In ulcerating acanthoma the base is formed of dense eroded tumor tissue; 
the secretion is scanty or abundant pus, serum and opaque particles of 
tumor-cells; the edges are raised, very hard, in mucous membranes often 
undermined, rounded, irregular, or serpiginous; and the vicinity is indurated 
and inflamed according to the extent of infiltration and infection. Erosion 
of bone and cartilage with pain, and of blood-vessels with hemorrhage may 
occur, and perforation into adjoining hollow viscera. Considerable masses 
may slough, and new skin or mucous membrane may replace portions of the 

Fig. 2. — Fungating carcinoma of neck. A recorrence from the lip. 

ulcer. Extensive tracts of suppuration may follow invasion of cellular 

In adenocarcinoma of mucous membranes thrombosis and infection com- 
monly leads to the formation of large sloughing excoriated ulcers with much 
hemorrhage. Flat, indurated, annular ulcers with extensive infiltration 
of the vicinity and constriction of the lumen form in carcinoma of the rectum. 
Necrosis of large portions of the tumor together with the distended overlying 
skin in carcinoma of the breast and other cancers produces extensive ulcers 
in which the sloughing tumor may protrude or suffer deep excavation. In 
sarcoma, ulceration is usually marked by massive necrosis and hemorrhage. 

The onset of ulceration frequently transforms a comparatively harmless 
tumor into a rapidly fatal process through local or general infection, suppura- 
tion, absorption of toxic products and hemorrhage, to which chiefly must 
be attributed the cachexia of cancer. 

In general it may be said that the outward morphology of tumors con- 



stitutes an important and very practical chapter in their study since it 
not only reveals the general architecture of neoplasms but allows their 
separation in most cases from other processes, and usually permits an accu- 
rate diagnosis of their histological structure. 


The study of tumor growth must begin with its earliest stages. 

In one large group this study leads to the scrutiny of misplaced a,nd 
chiefly embryonal cells, or even to the consideration of single cells, giving 
rise to the large group of tumors to which Cohnheim's theory applies. In 
another equally Important class the initial changes affect cells which so far 

Fig. 3. — Precancerous changes in breast. Atypical epithelial proliferation in ducts and 
acini found in a small portion of the breast. 

as we know are normal and here a series of preliminary changes occurs, 
belonging to the so-called precancerous stage, which is of importance in the 
origin and growth of tumors. 

Precancerous Stage of Tumor Growth. — Processes occasionally followed 
by neoplastic growth are observed in syphilitic leukoplakia, tuberculosis, 
Hodgkin's granuloma, chronic eczema, chronic mastitis, ulcer of the stomach, 
cirrhosis of liver, x-ray dermatitis, and the chronic processes associated with 
mechanical, chemical and thermal irritation. 

It is impossible to analyze all the factors acting in such processes. Upon 
their interpretation have been built many theories of the nature of tumor 
growth. Yet it is possible to enumerate several very constant morphological 
changes which in certain combinations seem to be most effective in exciting 
the subsequent autonomous growth. 



In a study of precancerous changes in several situations Bonney finds 
constant loss of elastic tissue, and usually hyaline changes in the collagen, 
lymphocytic exudate, and fraying of the edges of the epithelium. 

In multiple epithelioma of the skin Janeway has observed that the tumors 
are often preceded by a period of local erythema which may slowly disappear 
or be followed by cancer. Sections of the skin show focal dilatation of 
the blood capillaries adjacent to the basal epithelium about which the epithe- 
lium is hypertrophied, hyperplastic and with hyperchromatic nuclei. 

In leukoplakia, hyperkeratosis, thickening of the Malpighian layer, 
hypertrophy of tlje epithehum, round-cell infiltration, increased vascularity 
and edema long precede epithelioma of the tongue. 

A-Asste? Y 

Fig. 4. — Atypical epithelial hyperplasia on edge of a beginning carcinoma of skin. 

In ulcer of the stomach disordered gland alveoli are drawn into new cellu- 
lar connective tissue, detached from their normal relations, pass through 
stages of proliferation, hypertrophy of cells and nuclei, and gradually or 
suddenly assume neoplastic characters. Here, as in cancer following cir- 
rhosis of the Hver, it is sometimes difficult to determine just when a true neo- 
plastic process exists. 

In chronic mastitis single small cancer nodules may appear and I believe 
their preliminary stages may be traced in other portions of the breast, chiefly 
in the ducts. Here one finds hyperplasia of hning cells with hypertrophy 
and increased eosin stain, cyst formation and papillary outgrowths into them, 
in which condition they seem usually to be diverted from further dangerous 
tendencies. Yet other cells show more active proHferation, increase in size 
and in nuclear chromatin; their polarity is lost by growth of round-cells and 
connective tissue, and it is possible to trace such cells into types closely 
approaching those seen in early carcinoma. 


It may be true as Ribbert asserts,. that no one has ever seen the beginning 
of cancer of the breast, but the changes just enumerated seem to form a com- 
mon prehminary to the onset of true cancer, which on such a basis develops 
rapidly and is accompanied as a rule with fatty degeneration in the affected 

In rare cases of Hodgkin's granuloma the prohferation of large mono- 
nuclear cells and of giant-cells, originally of inflammatory origin, takes on more 
autonomous and sarcomatoid qualities. The cells become more numerous, 
the stroma less, the nuclei larger, mitoses more abundant, metastases com- 
posed of large round-cells without stroma occur, and the process has many of 
the qualities of a blastoma. Yet the true neoplastic properties of these cases 
must be doubted, although they exhibit a certain progress toward a genuine 
autonomous newgrowth. 

The abnormal physiology associated with the precancerous state will be 
discussed under General Etiology. 

Beginning thus either i;i isolated embryonal cell groups or in tissues 
long subjected to the influences of the precancerous stage, tumors grow, 
always chiefly, sometimes exclusively, from their own resources, i.e., from the 
original cells involved in the process. 

Modes of Origin. — Four modes of the origin and extension of tumors may 
be discerned: 

1. Growth is unicentric and exclusively from the original cell group 
involved in the tumor process. In the large class of blastomas arising from 
embryonal cell groups, the originating cells are isolated from the first, and the 
tumor-cells remain isolated during their entire history. Thus hyperneph- 
roma may begin in a small group of cells misplaced in the renal cortex, 
produce a large tumor of the kidney, invade the renal vein and produce 
metastases in many organs, the cells in all situations being the direct descend- 
ants of those composing the original cell complex. 

2. Multicentric growth is observed in many tumors. In multiple epithe- 
lioma of the skin, x-ray cancers, cancers of the gastro-intestinal tract 
following multiple polypi, fibrosarcoma of nerve trunks, and others, several 
centers originating tumor-cells appear simultaneously or at intervals. These 
centers may gradually coalesce producing one large tumor area or they may 
long remain discrete. 

Petersen by reconstruction in wax has shown that certain cancers of the 
breast arise from several discrete centers of prohferation, and a similar 
multiple origin in an adrenal tumor is described by WooUey; and for carcin- 
omatosis of the liver, by Von Heukelom and Oertel. 

A remarkable case of multiple angiosarcoma or endothehoma, involving 
nearly the entire skeleton, is described by Marckwald. 

Of such cases some must be referred to the wide diffusion of the inciting 
factors as in a;-ray cancers; others to multiple foci of embryonal displaced or 
superfluous cells as Hauser would explain the multiple cancers of the intes- 
tine ; while the conditions in a few cases suggest some complex influence of an 
existing tumor upon neighboring cell groups predisposed to tumor growth. 

Multiple tumors of the same type in paired organs is a special case of 
multicentric growth. Carcinoma arises in both breasts, teratoma in both 
ovaries, lymphosarcoma in both testes. In each such case different influ- 
ences seem to be concerned but in all the influence both of local and of con- 
stitutional factors is suggested. 

The occurrence of multiple tumors of different types in the same organ 
or in different organs of the same subject is also observed. 

In some instances a second tumor may result from conditions brought 


about by the presence of the first, as when a submucous fibromyoma causes 
ulceration followed by cancer of the uterine mucosa. 

Tumors have been observed simultaneously in the uterus, ovaries and 
breast, a combination suggesting the influence of the functional relation 
existing between these organs. The rather common occurrence of two or 
more tumors in different organs of the same subject suggests nothing more 
than the accidental coincidence in several organs of the general etiological 
factors in the genesis of tumors. 

Many cases illustrating several groups of multiple tumors have been 
collected by Walter and Woolley. Among looo autopsies on tumor cases 
Hansemann found five with multiple primary growths, and in 1225 cases 
collected by Redlich 14 showed more than one primary tumor. When a 
minute search is carefully conducted and all forms of tumor growth included, 
the proportion of multiple tumors is much greater. 

Symmers in 215 autopsies found tumors in 55, in 22 of which more than 
one type existed. Gotting observed three cancers of different organs, 
several adenomas of bile ducts, and an angioma of liver, in a man of 58 years. 

3. The question whether a tumor process may gradually extend from an 
originating focus over the surrounding normal cells cannot at present 
be determined. Such a possibility is commonly denied. Borst argues 
that it is impossible to conceive how such an influence can be transferred 
from one cell to another. Ribbert, after strongly opposing the idea of 
transference of the tumor tendency from a tumor-cell to normal cells, admits 
that there may be widespread multiple foci of origin which exhibit neoplastic 
growth one after the other, thus simulating the lateral diffusion of the tumor 
process. Unless one assumes as White actually does, that tumors develop 
from single cells, it is necessary to assume some such gradual implication of 
the several cells in the originating group, so that Ribbert has practically 
admitted the gradual extension of a tumor by progressive involvement of 
normal cells, not by transfer of influence from one cell to another, yet by 
the effects of underlying factors upon one cell after another. It is not 
necessary to attribute this influence to the mere contact of tumor-cells or 
to suppose that it involves a parasite. 

Increased vascularity, mechanical displacement, disordered function, 
chemical stimuli and toxic agents, may all be included among the factors 
resulting from the presence of tumor-cells. There is also a certain complex 
tropism observed in the growth of tissues in lower animals, whereby processes 
at work in one tissue induce similar tendencies in contiguous cells. There 
appears to be adequate theoretical basis for the extension of a tumor process 
over normal cells, and observation indicates that in some cases such extension 
actually occurs. 

First, in the neighborhood of many epithelial tumors the contiguous 
cells show a peculiar increase in number, size and chromatin content. 

This collateral hypertrophy is more common about primary tumors but 
may be seen in metastases. Usually it does not reach a neoplastic grade, 
but it is sometimes difficult to distinguish from an advance of the tumor. 

In adenoma of the sigmoid such transformations may occur, the outlying 
cells first increasing in size, then becoming opaque and granular, while the 
nuclei become hyperchromatic. Some extensive superficial adenomas of the 
sigmoid appear to have developed in this manner. 

Hauser depicts definite stages in the transformation of normal glands 
into hypertrophic forms which he believes precede the development of 
cancer. Haaland has described a neoplastic transformation of the pulmonary 
cells in metastatic carcinoma in the mouse, an observation which needs to 


be confirmed, but the sarcomatous transformation of the stroma of trans^ 
planted cancers of the mouse, if genuine, is probably a sufficiently attested 
instance of the tropism exerted by tumor epithelium upon adult connective 

In adenocarcinoma of the human ovary is sometimes seen a sarcomatoid 
proliferation of the stroma. 

Yet these instances of lateral extension of tumor processes, if they eventu- 
ally stand the test of criticism, are rare and it should be emphasized that 
the great majority of tumor-cells are isolated in origin and throughout their 

4. The systemic round-cell sarcomas present a difficult problem in 
the analysis of their mode of origin and extension. Multiple myeloma arises 
simultaneously at numerous widely distant foci in the bone-marrow and 
may early show a diffuse tumor process throughout many bones. Hence 
one must assume the existence of so many foci of origin as to suggest the 
simultaneous involvement of the entire bone-marrow system. 

Multiple metastases cannot satisfactorily explain the course of these 
tumors, and it would seem probable that there occurs a lateral extension 
of the tumor process over adjacent normal areas as well as infiltrative growth 
from the multiple foci. In the group of lymphocytomas there are examples 
very strongly indicating simultaneous tumor growth originating in a large 
portion of the lymphatic system. In some intestinal cases (Wells, Symmers) 
there is a diffuse lymphadenomatous hjrperplasia involving the entire lym- 
phoid apparatus, from the cardiac orifice to the anus, the superficial, the thor- 
acic and the abdominal nodes, and the bone-marrow. Leucocytosis is 
absent and the spleen is little affected. A rapid lateral extension from very 
many foci or a primary involvement of the entire system seems to be the 
only reasonable explanation of such cases. There are systemic highly 
malignant large-cell sarcomas of equal extent. Chloroma, with or without 
leukemia, occurs as a multiple or diffuse tumor process affecting wide areas 
of marrow and lymphatic system for which a lateral extension from many 
foci or a primary diffuse origin seems a necessary assumption. In thyroid 
tumors of fish the hyperplasia which seems to result from exaggerated re- 
sponse to functional stimulus, is felt in all the widely disseminated alveoli 
of the gland. Some of these hyperplastic glands may regress (Marine), 
but beyond a certain stage they become autonomous. 

Cell Division in Tumors. — Normal and many abnormal types of cell 
division occur in tumors, including mitosis, direct nuclear division, and 
direct and indirect fragmentation. 

This subject has been very fully investigated by Stroebe, Vit. Miiller, 
Cornil, Arnold, Pianese, Galeotti, Hansemann, and many others, who agree 
that the chief mode of cell division in tumors is by mitosis. In the benign 
tumors in which the growth is comparatively slow the scanty mitotic figures 
usually appear normal, but degeneration, inflammation and very rapid 
growth are marked by increasing irregularity in the mitotic process. 

Pianese classifies the irregularities occurring in cancer as follows: 

A. Relatively typical mitosis, which may be bipolar and symmetrical, 
hyper- or hypochromatic, and giant. 

B. Atypical mitosis, either asymmetrical or multipolar. 

C. Abortive mitoses, marked by suppressed or atypical polarization of 
chromosomes, by aberratioti of chromosomes, by loss of polar bodies, and by 
degeneration of cytoplasm. 

As a rule he found the first two types in comparatively intact cells, while 
abortive mitoses occurred chiefly in degenerating cells. Irregularities in 


all the phases and elements of the mitotic process occur in malignant tumors 
and bear a certain relation to the morphological and physiological variation 
of the tissue from the normal (Hansemann). The number and size of the 
chromosomes, the structure of the spindles, the coordination of the division 
phases, the position of the elements in the cell, contribute numerous abnor- 
malities in the process. Hyperchromatic cells may arise from asymmetric 
mitosis. Hypochromatism results from asymmetric mitosis and loss of 

Asymmetric mitosis has been referred to secondary subdivision of cen- 
trosomes, one of which may divide into as many as four parts, each forming 
attraction spheres. Under these conditions the migration of chromosomes 
is often delayed and unequal (Galeotti). 

The results of abortive mitosis give rise to a great variety of structures 
which have been elaborately illustrated by Pianese. In this class of struc- 
tures have been found many of the pseudoparasites of cancer. Thus abnor- 
mahties in the mitotic process have been observed rather more abundantly 
in carcinoma than in sarcoma, but many of them have been seen in inflam- 
matory and regenerative hyperplasia (Stroebe). 

Striking pecuHarities in such a fundamental process as cell division 
may well suggest that here is disclosed the essential nature of tumor growth. 
Hansemann has interpreted the studies in this field in such a manner, 
holding that asymmetrical, multipolar and abortive mitoses mean unequal 
distribution of cell potencies, loss of cell differentiation, freedom from normal 
restraints to growth, and exaggeration of growth over functional capacities. 
These physiological properties he includes in the conception of anaplasia, 
the morphological side of which is illustrated in resting and dividing tumor- 
cells. To a considerable extent this interpretation may be valid, and yet 
the majority of tumor-cells divide by normal mitosis (Pfitzner, Hauser), 
and the minute study of the conditions favoring abnormal mitoses indicates 
that these changes are secondary results of tumor growth and not primary 
and essential. 

Abnormal, asymmetric and multipolar mitoses are readily produced 
experimentally, by heat, and by many chemicals (Galeotti, O. Hertwig). 

Gametoid Mitosis. — In the maturation of sex cells the mitotic nucleus 
exhibits only one-half the number of chromosomes of the somatic or general 
body cells, and these chromosomes instead of assuming a V shape and radial 
arrangement and splitting lengthwise in the monaster, are ring-hke or loop 
shaped, or composed of coarse granules arranged in the long axis of the 
spindle. This type of division with reduction of chromosomes is called 
gametogenous or heterotype. 

Farmer, Moore and Walker have found that this type of division, here 
called gametoid, is common in the growing edges of epithelioma and other 
tumors. On this ground they at first assumed that tumor-cells possess the 
physiological significance of sexual cells, but this view has been abandoned 
since gametoid mitosis is by no means constant in tumors, and the pecuhar 
mitoses may not signify the properties of sex cells. Yet, I think the obser- 
vation remains suggestive, since tumor-cells may show three properties of 
sex cells, gametoid mitosis, increased growth capacities, and remarkable 
altruistic relations. 

Amitosis is frequently observed in rapidly growing tumors and in some 
cases it may be the chief mode of cell division. According to Nedjelsky 
it begins with swelling of the nucleolus followed by elongation, budding or 
cleavage of the nucleus. It results in the formation of giant-cells with pale 
nuclei which may later undergo budding or mitotic division, or by division 


of the cell body it often leads to rapid regeneration of viable tumor-cells 
(V. Miiller). 

Flemming and Trambusti hold that amitosis always results from degenera- 
tive changes in the cell and may yield giant-cells but not fully viable tumor- 
cells. Marchand and most observers believe that mitosis contributes, 
as does amitosis, to the rapid regeneration of tumor-cells, but fails to convey 
to the progeny all the properties of the parent cells. Pianese in his elaborate 
study of cancer-cells could not find that amitosis was ever followed by cell 
division, but was always abortive, leading to the formation of multinucleated 

Howard has endeavored to trace in the nuclear changes of tumor-cells 
many processes parallel to those observed in the protozoon nucleus. 

The impression that tumor-cells must owe their exaggerated growth 
capacities to some form of fertihzation has led to an interpretation of certain 
nuclear changes suggested by this point of view but now abandoned. Klebs 
once thought that the nuclei of englobed leucocytes fused with the nuclei of 
tumor-cells. Recklinghausen saw appearances suggesting the fusion of 
the nuclei of endothelium and fibroblast. 

Auerbach and Bashford pictured the fusion of the nuclei of adjoining 
tumor-cells, and Schleich thought that endogenous cell formation consti- 
tuted a form of cell infection which accounted for overgrowth. 

Neoplastic vs. Inflammatory Hyperplasia. — Neoplastic hyperplasia is 
usually more rapid than the multiplication of cells in inflammation, or re- 
generative growth. Lymphosarcoma of a few months' growth may reach 
a bulk constituting a large portion of the bodyweight. Carcinoma of the 
hver may largely replace the tissue of this organ in a brief period. In the 
terminal stages of many malignant tumors growth is much accelerated so 
as to divert much nutriment from the other tissues. This rapid growth is 
associated with a great abundance of mitotic nuclei in the tumor-cells which 
often far exceeds that seen in any inflammatory process. 

Yet rapidity of growth is not an essential distinguishing feature of 
tumor processes. Few tumors grow as rapidly as the fetus in utero. 

Leucocytes' multiply with great rapidity in pneumonia; and epithehum 
covers denuded skin far more actively than the growth of such ceUs in many 

A more important distinction of tumor growth is found in its progressive 
tendencies. Tumors commonly grow to a bulk far exceeding the Hmits 
of any inflammatory or physiological process affecting the same area, and 
they continue to grow to whatever extent nutrition is provided. Physiolog- 
ical regeneration observes a certain adaptation to the function of the 
organ; inflammatory overgrowth is measured by the intensity and duration 
of certain irritants and ceases when these irritants, are removed; tumor 
growth ignores all these hmitations and ceases only with the death of the 
patient or the curtailment of blood-supply. Yet there are cyclic variations 
in the growth of tumors, periods of activity being followed by standstill or 

Indeed there appears to be a natural Kmit to the growth of some tumors 
in some subjects. Progressive fibrosis brings some cancers of the breast 
to comparative quiescence in an innocuous scirrhus form, and the meno- 
pause marks the acme of growth of some uterine myomas, and has been 
followed by the spontaneous regression even of cancer of the breast (Watson- 

The frequency of tumor regression in lower animals, the increasing number 
of spontaneous regressions in human] tumors (Czerny), and observations of 



the unhindered growth of tumors over many years in old persons, indicate 
that limitless growth is not a power possessed by all mahgnant tumors. The 
universal employment of surgical measures renders this fact less obvious 
than it might be. 

The morphology of neoplastic hyperplasia differs from inflammatory 
overgrowth and physiological regeneration. A productive inflammation 
is usually a diffuse process involving a whole organ or tissue area and it grades 
off insensibly into normal tissue (chronic mastitis), while a neoplasm is 
usually circumscribed on account of its origin from an isolated group of cells 
(fibroma mammas). Productive inflammation involves all the elements of a 
tissue; tumor processes usually reside in a single-cell type. Inflammatory 
neoplasia is confined by narrow limits, which in tumors are very wide. 

Fig. S. — Mammary cancer. Spontaneous local regression, with formation of fibrous tissue, 
calcific granules, and fatty crystals. 

Tumor-cells show the features of anaplasia, loss of specific form, and of 
polarity, variation in size, nuclear hyperchromatism, multiformed nuclei 
and atypical mitoses, all of which are absent or less marked in inflammatory 

Yet inflammatory overgrowth passes by insensible gradations into neo- 
plastic hyperplasia. This significant principle of pathology has an important 
and yet a limited application. As a rule tumor growth differs sharply from 
any inflammatory h3^erplasia affecting the same tissue, but there are excep- 
tions to this rule. 

(a) There are inflammatory hyperplasias of the bone-marrow cells, 
leukemic processes with some of the features of a neoplasm, and chloromas 
which are true tumors of marrow-cells. Most cases fall clearly into one 


of these categories but others are intermediate in type and it is difficult to 
determine whether they are strictly inflammatory or truly neoplastic. 
Similar intermediate grades of hyperplasia, difficult to classify, are occasion- 
ally encountered with nearly every tissue cell in the body. 

(h) A process beginning as a simple inflammatory hyperplasia may in 
the same individual gradually assume neoplastic properties. 

In the thyroid gland of goiter, especially in fish, in the prostate gland of 
hypertrophic prostatitis, in the uterine mucosa of a glandular endometritis, 
in the mammary gland of chronic mastitis, and in the lymph-nodes of Hodg- 
kin's granuloma, are occasionally seen transformations of a functional or 
inflammatory hyperplasia into a more or less typical neoplasm. Here it 
would appear that a hyperplastic process incited by external irritants may 
continue and increase long after the removal of the irritant. As with certain 
chemical reactions so with pathological processes the importance of momen- 
tum must be recognized. 

It may be objected that when a neoplasm appears in the course of in- 
flammatory hyperplasia something new is added. 

This objection may be valid for certain cases where the neoplasm appears 
suddenly and in widely diff^erent form from that exhibited by the preexisting 
hyperplasia, but not for other cases in which observation shows the change 
to be gradual. Since with ever)' increase in quantity there is a change in 
quality it is reasonable to expect that gradually accumulating tendencies 
in the life of the cell may express themselves at times in quite sudden and 
radical changes in morphology and behavior. All these considerations 
strengthen the view that the "precancerous condition" is of wide occurrence 
and of much theoretical and practical importance in oncology. In this condi- 
tion one finds tissues and cells in a state of overgrowth intermediate between 
inflammatory and neoplastic hyperplasia, exhibiting certain tumor characters 
which must be judged from different standards for each tissue, and which 
experience shows are often followed by genuine .and usually malignant 

The theoretical distinctions which a general survey establishes between 
neoplastic and inflammatory hyperplasia are sharp and fundamental but 
these distinctions fail us when we have to search for them in processes of 
doubtful nature. Here we assume them to exist from our general knowledge 
but we cannot prove their presence. 

Thus as Borst assumes, inflammatory overgrowth results from exagger- 
ated response to external irritants while neoplastic growth arises from loss of 
normal restraints to growth. The one is purposeful, organized, self-limiting, 
typical, accelerated hyperplasia, while the other is baneful, lawless, pro- 
gressive, atypical degenerative growth excess. But this is a composite 
picture and the practical problem is to decide how much it is reflected in 
individual doubtful cases. Virchow recognized this difficulty and some ob- 
servers, finding it insurmountable, have concluded that there is only a differ- 
ence in degree bet\Veen heahng of wounds, inflammatory hyperplasia and tumor 
formation (Brosch, Fabian). 

This view cannot be accepted. Inflammatory hyperplastic processes 
pass by insensible gradation into neoplastic growth according to our present 
analysis. SyphiHtic leukoplakia becomes epithelioma, and Hodgkin's 
granuloma may take on a sarcomatoid character. Yet these are complex 
processes and although we cannot detect when growth from external stimula- 
tion is succeeded by growth from loss of growth restraints, yet theoretical 
considerations require us to assume that at some point such a change actually 
occurs. I have elsewhere suggested that some hyperplastic processes, owing 


their inception to the influence of parasites, become at length established 
and progress without the stimulus of the parasite. Here one may suppose 
that the declining stimulus is replaced by progressive loss of growth 
restraints, but the reasoning would still show that inflammatory hyperplasia 
passes into neoplastic growth. 

On the other hand I do not beheve that growth restraints are fully main- 
tained in some inflammatory hyperplasias. In proud flesh mechanical 
restraints seem obviously defective and in hypertrophic endometritis several 
signs of lowered tissue tension are visible. External stimulus and loss of 
growth restraints are here combined, and it cannot be said that neoplastic 
hyperplasia is wholly different in nature from inflammatory overgrowth. 

Thus, however closely the processes are analyzed, the conclusion remains 
that inflammatory hyperplasia passes into neoplastic. 

Regressive Changes in Tumor-cells. — Tumor-cells in general suffer the 
same degenerative changes as occur in other tissues but some special interest 
attaches to this field, on account of the frequency and extent of these changes 
in tumors, and also from the very minute analysis which has been required 
because of the former interpretation of many intracellular structures as 
parasites. The presence of degenerative changes in tumors is the direct result 
of the activity of growth, the accelerated metabolism, and the uncertainty 
of blood-supply. 

Fatty degeneration is nearly constant in cancer, less common in sarcoma. 
It contributes a light opaque color to tumor cords which is one of the earUest 
gross diagnostic signs of many cancers. It seldom reaches an extreme degree, 
the fat being commonly limited to fine droplets which first appear about the 

It is most prominent in tumors of glands which normally secrete fatty 
substances as in adenoma sebaceum, adrenal tumors, and cancer of the 
breast. In rapidly growing, well-nourished tumors the active metabolism 
is unfavorable to the deposit of visible fat, which appears chiefly in areas of 

The chemical nature of the lipoids of tumors is no better known than is 
that of the fats in other conditions. In a series of cancers Pianese could find 
no fat blackening with osmic acid. 

Stains by Sudan III and Scharlach R commonly reveal many fine fat 
droplets in such tumors. To what extent the fatty degeneration of tumors 
belongs in the class of lipoid or myelinic degenerations has not been deter- 
mined, but Kaiserling and Orgler found myelin droplets in the cells of many 
cancers and sarcomas. 

White has described several varieties of fatty crystals occurring in frozen 
sections in and about the cells of carcinoma, sarcoma, and adenoma. Some 
of these were large plates or needles which did not melt readily and which 
he regarded as cholesterin. 

Smaller needles which melt and solidify into globules on cooling he identi- 
fied as fats. Other small needles and anisotropic crystals were classed as 
mixtures of cholesterin with lecithin and as myehn forms. The crystals 
seem to be associated with cell-growth rather than with degeneration and 
White suggests that they are in some way concerned in the regulation of 

Hydropic degeneration occurs in the edematous portions of both benign 
and malignant tumors, and affects both nuclei and ceU bodies. 

It is seen in marked degree in the fining cells of cystic adenoma and iri the 
cells of any tumor that is constantly bathed in fluids as in papillary epithe- 
fioma of the bladder. ExfoHated cells lying in tissue spaces are commonly 


ballooned out with imbibed fluids. Virchow's "physaHden" were in part 
hydropic vacuoles in nuclei or cytoplasm which in the fresh condition 
appeared hyaline but on hardening yielded a central granular or homogene- 
ous mass. 

Glycogenic degeneration of tumors has been carefully studied by several 
observers. According to Gierke it appears in tumors under the same condi- 
tions as in other cells. These are mainly as a primary physiological con- 
stituent of actively growing or embryonal cells and as a secondary result of 
disturbed nutrition and degeneration. In the former class it would naturally 
characterize malignant, rapidly growing and embryonal tumors, and in the 
latter it would be associated with and have much the same significance as 
fatty changes. Observation has shown that glycogen is commonly found in 
most rapidly growing tumors (Best), but in very variable quantities, and 
some malignant growths as breast cancer are usually free from it. In endothe- 
lioma it is very constant and abundant (Gierke). In 1544 tumors Lubarsch 
found glycogen in 477, 29 per cent. Fibroma, osteoma, glioma, and heman- 
gioma were constantly glycogen-free; lipoma and lymphangioma and ade- 
noma nearly always so. The more certain the embryonal nature of the 
tumor the more constant was the presence of glycogen. Thus in teratomas, 
h)rpernephromas, and choriomas it was present in all cases and usually 
abundant. It was about equally common in sarcomas (50.7 per cent.) and 
carcinomas (43.6 per cent.); in squamous cell epithelioma 70 per cent. 

Four factors seemed to Lubarsch to chiefly determine its occurrence. 

1. An origin of the tumor from embryonal cells. 

2. An origin from cells normally containing glycogen. 

3. The absence of mucous and colloid degeneration, and frequent presence 
of fatty changes. 

4. The presence of numerous delicate blood-vessels in intimate relation 
to the parenchyma. 

Secondary glycogen deposits appear in necrotic areas and about the 
inflamed edges of tumors which are elsewhere free from it (Best). Brault 
thought that glycogenic degeneration sometimes preceded mucous changes 
in mucoid carcinoma. 

Miiller concluded that actively growing, well-nourished cells contain 
no glycogen, which appears in cells with feeble metabohsm. The relation 
of glycogen content to the prognosis of tumors was considered by Brault, 
who states that in sarcomas there is no parallel between glycogen content 
and malignancy, while with carcinoma the glycogen holding tumors are 
regularly the more malignant. Best and Gierke, however, did not find this 
rule to hold. 

Mucous degeneration is a widespread and important regressive change 
in many tumors. It may be a primary and essential feature of the neoplasms, 
arising from exaggerated development of the mucous constituents or from 
increased secretion, or it may overtake tumors whose cells have no original 
tendency toward its production. 

It represents the matrix of primary myxoma, the secretion of many 
adenomas, and a degenerative product in chondroma, sarcoma, and carci- 
noma. It is readily recognizable by the homogeneous luster and viscid 
quality which it imparts to the tissue. True mucin is found chiefly in 
connective tissue tumors, pseudomucin in epithelial growths. Certain 
tumors are characterized by a remarkable production of mucus. Primary 
myxoma of the skin consists of spindle- and star-shaped cells lying in a 
matrix of homogeneous finely fibrillated material which infiltrates the sur- 
rounding tissues and doubtless contributes to the local recurrence of these 


growths. Some gelatinous cancers of the colon early invade the peritoneum, 
spread through the cavity and produce immense quantities of firm gelatinous 
material which distends the abdomen. In such cases it may be difficult 
to discover any remnants of living tumor-cells, and some such tumors seem 
to reach a standstill through excessive production of mucus. Cancer of the 
stomach may yield a localized growth or diffuse infiltration of the stomach 
wall in which are found scanty groups of tumor-cells widely distended by 

Adenoma of the ovary may yield cystic tumors of very large dimensions 
composed of cysts distended by fluid containing large quantities of pseudo- 
mucin. Mixed chondroma of the parotid commonly shows areas of cartilage 
verging into myxoma, and recurrences of the tumor may eventually assume 
the type of pure myxoma. Actively growing lipomas are occasionally over- 
taken by diffuse mucoid degeneration. 

Intracellular mucous globules are observed in a wide variety of tumors 
in which the process fails to make headway. As a rule, when mucous 
degeneration overtakes a tumor, while a rapid increase in size may follow, 
the tumor has reached the acme of its cellular activity and its malignancy 
is reduced. Since the nuclear constituents of the cell are chiefly concerned 
in mucus production, the appearance of intracellular mucous degeneration 
usually indicates a process intimately affecting the vitality of the cell. 

Hyaline changes in cells, vessels and stroma, are the most common forms 
of regressive process occurring in tumors. Firm, homogeneous, acidophile 
material arises from the connective tissue elements and vessels of the stroma, 
from the tumor-cell cytoplasm, and from exudates and secretions. Pro- 
nounced grades of the first type are seen in keloids, fibroma, fibrocarcinoma, 
endothelioma, and angiosarcoma. 

It may strongly influence the course of the neoplasm, usually appearing 
in slowly growing or regressing tumors to which qualities it contributes. 
Atrophy of cells usually keeps pace with the progress of hyaline changes, 
so that large tumors may be converted into nearly stationary growths com- 
posed largely of hyaline masses. 

In some angiosarcomas the walls of vessels are converted into hyaline 
rings sheathed by ill-nourished tumor-cells. Calcification of these hyaUne 
vessels produces the sand grains of psammoma. 

Intracellular hyalinosis produces a wide variety of intracellular structures 
which have been studied and elaborately depicted by Pianese. A some- 
what special variety is the keratinization of squamous cell epitheliomas, 
in which group the appearance of hyaline globules, rings, and vacuolated 
bodies, within or about the nucleus, in the cytoplasm, or involving the cell 
borders, produces bizarre structures, many of which were once interpreted 
as parasites. Exudates and secretions becoming inspissated contribute a 
large portion of the hyaline material in some tumors. "Mulberry cells" 
containing small nuclei and many large acidophile globules resembling red 
blood-cells occur in certain tumors and in other conditions. Their appear- 
ance suggests an origin from englobed red cells or from intracellular hyahne 
degeneration. Weber depicts them in a case of myeloma, and I have found 
them very abundant in cancer of the stomach. True amyloidosis has been 
described by many authors in a variety of tumors, as carcinoma of esophagus 
by Wagner, tumor of the bladder (Billroth), carcinoma and sarcoma of 
larynx (Burow), carcinoma of lung (Langhans), chondroma of lung (Siegert), 
carcinoma of breast (Aoyama) and endosteal sarcoma (Hildebrandt) . Some 
fibromas or sarcomas of the larynx are characterized by the development of 


large masses of amyloid. Usually the amyloid masses have a concentric or 
radial striation. 

Colloid Degeneration. — Homogeneous, semisolid, acidophile material occurs 
in many tumors in large intercellular masses and in small intracellular 
globules. The largest collections are seen in ovarian adenomas, in which the 
ordinary mucin may present the qualities of iirm colloid. Thyroid tera- 
tomas containing little else than thyroid tissue with distended alveoli are 
frequent in the ovary, and some ovarian adenomas, not of thyroidal nature, 
may show considerable masses of colloid substance. Carcinomas in various 
situations may present cylindrical intra-alveolar masses of colloid approach- 
ing the appearance of thyroid tissue. Colloid appears in the cells of many 
tumors in the form of globules of rather dense strongly acidophile material. 
These structures are one of the chief sources of Russell's fuchsin bodies. 
Colloid droplets may appear in well-nourished mitotic cells or in cells showing 
advanced stages of other forms of degeneration. 

Miscellaneous Peculiarities of Tumor-cells. — Besides the well-defined 
types of degeneration tumor-cells exhibit a wide variety of alterations which 
have been subjected to close scrutiny but which it is difficult to interpret 
and classify. 

These changes affect both nucleus. and cytoplasm. 

In the highly developed tumor-cell nucleus may be distinguished five 
elements : 

1. Nuclein, a nucleoproteid, composing the chromatin which appears 
in densely basic staining clumps, normally arranged along the periphery of 
the nucleus or as an intranuclear network. 

2. Paranuclein, an acidophile substance composing the nucleoli. 

3. Linin, or plastin, an achromatic substance forming an intranuclear 

4. Amphipyrenin, which some authors describe as constituting the 
nuclear membrane. It is distinguishable from linin. According to Albrecht 
the nuclear membrane is a' lipoid substance. 

5. Nuclear fluid, filling the meshes of the nuclear network. 

As compared 'with normal cells all of these nuclear structures may be 
much more abundant in tumor-cells. In giant-cells especially those of 
myogenous origin, and in various sarcomas and epitheliomas, the nuclei 
reach astonishing proportions from hypertrophy chiefly of nuclein. In 
epitheliomas great excess and multipHcity of paranuclein bodies is somewhat 
characteristic, while in endotheliomas the nucleoli are relatively small. 
Shrinkage and pyknosis of nuclei occurs in degenerating and necrosing areas 
of many tumors. Wide distention from imbibition of fluids may accompany 
hydropic degeneration of the cytoplasm. Rarefaction of nuclein masses 
may result in a pale diffuse stain of the thickened chromatin network. Or 
the chromatin may appear in very thin strands eventually disappearing in 
complete karyolysis. A common appearance in carcinomas is the presence 
of several discrete blocks of chromatin lying irregularly in a nucleus devoid of 
chromatic membrane. The various stages of karyorrhexis may be followed 
in degenerating or necrosing cells. The fragments of chromatin thus result- 
ing may be scattered in the cytoplasm and remain pyknotic or become 
vacuolated or surrounded by cytoplasmic vacuoles or become dissolved. 
Extrusion of chromatin into the cytoplasm may result in the appearance of 
many basophile granules in the cell. The chromatin or linin may break up 
into many fine rings within the nucleus and these be discharged into the 
cytoplasm, as observed by SchuUer and interpreted by him as the forms of an 
intranuclear parasite. 



Under the term pseudo-adipose and cystic degeneration of nucleus 
Pianese described the appearance of homogeneous spheroidal, brownish, 
achromatic globules in the nucleoplasm, occurring especially in tumors of 
the liver. This author also describes a hyalinosis of the nucleoplasm which 
results in a transformation of part or the whole nucleus into a hyaline mass. 

Fragmentation . of the nucleus with formation of pseudonuclei occurs 
in many giant-cell tumors. In melanoma the nuclei often contain large 
masses of homogeneous material in which appear all stages of the formation 
of granular pigment. 

In the cytoplasm of tumor-cells the early stages of degeneration, hydropic, 
mucous, lipoid, fatty, glycogenic, colloid, hyaline and calcific yield a great 
variety of distortions of the cell structure and abnormal granules, rings, 
globules and complex masses most of which are identified by appropriate 
technics. The complex morphology of these products is illustrated in the 
extensive literature on the pseudoparasites of cancer, especially in the work 
by Pianese. Many of the bodies, especially those of epithelioma, I have found 
duplicated by diphtheria toxin necrosis of the corneal epithelium. 

Fig. 6. — Bird's eye inclusions in cancer-cells. (Afler Le Count.) 

Phagocytosis by tumor-cells yields many intracellular bodies as from 
red blood-cells, leukocytes, other tumor-cells, bacteria, pigment, and foreign 
bodies, all in stages of degeneration. Meser found throughout a deep cancer 
of the breast lycopodium seeds which had penetrated the tumor from the 
dressing and had been englobed in the cells. 

Multiplication of centrosomes with their surrourlding archoplasm may 
yield a number of cytoplasmic bodies containing a central chromatic point, 
and a rim of condensed protoplasm more or less separated from the remaining 
cytoplasm. Hansemann found as many as twenty centrosomes in one can- 
cer-cell with multipolar mitosis. In normal cells the occurrence of multiple 
centrosomes has been abundantly shown by many histologists; Heidenhain 
counting as many as one hundred in a giant-cell of the rabbit's marrow. 

Comparing the centrosomes of the guinea-pig's spermatocytes with 
structures in cancer-cells, Borrel has shown that multiple centrosomes are a 
prolific source of the pecuhar bird's-eye inclusion of Leyden which he and 
many others have regarded as a parasite. He finds that the centrosomes of 
cancer-cells may appear as chains or congeries of chromatin granules without 
special protoplasmic mantle or as multiple chromatic granules with well 


segregated and somewhat differentiated cytoplasm, or as single swollen 
chromatin bodies with large cytoplasmic masses lying in cell vacuoles. 

LeCount's observations support Borrel's interpretation of a large class 
of intracellular bodies in cancer. Benda in dissenting from this view was 
able to show the presence of normal centrosomes in cells containing the bird's- 
eye inclusion, possibly from some other source than centrosomes. 

Accumulation of secretory products greatly modifies the appearance of 
many tumor-cells and produces some structures which have been interpreted 
as parasites. Since the secretory activities of glandular epitheUum are 
closely associated with that portion of the cell containing the centrosome, the 
presence of chromatin in many cancer bodies- is perhaps not an argument 
against the origin of these bodies from secretory processes. That such an 
origin accounts for many of the peculiar structures in cancer-cells has been 
maintained by many observers. Pianese thus explained the majority oi 
the bodies of the Thoma-Sjobing type which were small perinuclear bodies 
with relatively large chromatin granules. According to Greenough the 
characteristic bird's-eye inclusion of cancer-cells is practically limited to 
glandular carcinomas and is chiefly the result of secretory processes. 

He found them nearly constantly in breast cancers, but absent in sarcomas 
and epitheliomas. 

Somewhat similar conclusions have been reached by Nosske, Honda and 
Klimenko. As a rule these authors found the typical inclusions in slowly 
growing cancers in which secretory function may be assumed to persist. 
In position they usually lie between the nucleus and the alveolar lumen 
(Greenough) . They vary in size from minute globules to masses distending 
the cell. Stains show a central basophile portion which may be of large size 
or may be absent, and a rim of protoplasm which may be granular, or radially 
striated, or surrounded by a condensation capsule. While PUmmer stated 
that these bodies occur in enormous numbers in rapidly growing cancers, 
chiefly in the growing edges, later observers have not confirmed this state- 
ment but have found them in moderate numbers, chiefly in slowly growing 
tumors, and well within the edges. Apolant and Embden noted very few 
inclusions in mitotic cells, an observation which accords equally well with 
the secretory origin as with the centrosome theory. 

Thus the specific bird's-eye inclusions of cancer-cells, first described by 
Virchow, suggested by Foa as a possible parasite, appropriated and variously 
described by Thoma, Soudakewitsch, Leyden, Plimmer, Ruffer and Walker, 
and finally defended by Feinberg, must be ascribed to two main sources, 
multiplication of centrosomes and secretory processes in which centrosomes 
are probably concerned. The chromatin element of these bodies favors their 
origin from centrosomes, while the location in the cell and the morphology 
and occurrence chiefly in glandular carcinoma support the theory of retained 


The composition of tumors and the characteristics of their growth present 
many variations from the normal standards. 

Some tumors, called histioid, reproduce simple tissues and are composed 
of a single and uniform parenchyma. 

Others, called organoid, are of more complex architecture consisting of 
tumor-cells supported by normal stroma and blood-vessels, and reproduce 
the structure and preserve more or less the function of organs. The extremes 
of these types are well defined, as lymphosarcoma versus adrenal adenoma, 
but the great majority of tumors reveal some organoid characters. 



Homotypic or homologous tumors reproduce the structure of the tissue 
in which they arise (adenoma of breast) ; heterotypic or heterologous growths 
differ in structure from the tissue in which they grow (chondroma of breast). 
Tumors may reproduce the structures of adult organs, and tumor cells 
may preserve the characters of adult cells, as in epitheUoma of the lip, and 
these may be termed adult tumors. Others arise from groups of embryonal 
cells whose adult characters they never reproduce and these are conveniently 
termed embryonal. Examples are the socalled fetal adenoma of thyroid, 
and many teratomas of testicle and kidney. Such embryonal tumors are 
to be distinguished from those which, because of extreme metaplasia, differ 
widely from the originating structure. The organoid homotypic and adult 
characters tend to disappear as growth proceeds more rapidly and in meta- 
stases and recurrences, or in different parts of the same tumor. In ordinary 
cancer of the breast there may be wide variations in these quahties. 

Fig. 7. — PerithelioTnatous structure assumed by adenocarcinoma of uterus following 


Adamantinoma in successive recurrences may change from an adult 
acanthoma through an adenomatoid growth to a highly malignant tumor 
composed of closely packed, indifferent round and polyhedral cells. 

Chondro-epithelioma of the parotid of comparatively adult type may recur 
as pure atypical myxoma. Recurrences after operation may differ widely 
from the original structure. I have observed typical perithelioma after 
the removal of adenocarcinoma of testis, and in a uterus removed shortly 
after curettage for typical adenoma malignum I have found a spindle-cell 
tumor of perithelial type. 

Wholly spontaneous changes are illustrated in adenocarcinoma of adrenal 
which may invade the liver under the form of diffusely growing large cell 

In normal growth cells maintain a certain quantitative relation to the 
other elements of the tissue of origin, and they have an orderly arrangement 


and a polarity toward each other and to contiguous structures, but in tumors 
there is a progressive loss of these relations. Even in benign tumors as- in 
chondroma, the cells are in excess over stroma and their relation and position 
are altered. 

Some adenomas, as of thyroid and stomach, show a remarkable preserva- 
tion of normal structure even while proving malignant, but in most pro- 
nounced blastomas cells increase enormously over stroma, eventually 
relations between individual cells and between cells and stroma are lost, 
and the cells grow diffusely. Several notable characters of tumors are 
dependent upon the gradual development of these tendencies. Benign 
tumors retain much of the tissue organization, grow from their own resources, 
provide themselves with their own stroma and vessels, and push surrounding 
tissues before them and remain encapsulated. 

In malignant tumors the cells increase in number and size, become dis- 
placed from natural connections, lose their polarity and invade and destroy 
surrounding tissues. 

Lining cells grow outward into polypoid projections, drawing blood- 
vessels with them, or downward forming compressed pearls or convoluted 
cords; gland alveoli elongate and widen, or sacculated projections from sides 
or fundi produce simple tubules, or cystic or complex adenomas; diffuse 
masses of cancer or sarcoma cells disorganize the normal structure and force 
their way into neighboring spaces, alveoli and vessels. The demands for 
nourishment lead to formation of elongated vessels sheathed with tumor- 
cells producing the arrangement of perithelioma, which may appear in many 
varieties of tumors where the cells are pressed for food. 

Necroses arise where nutrition fails. Finally, the disordered cells pene- 
trating lymph- and blood-vessels, 'become loosened and pass into the circula- 
tion, lodging into adjacent lymph-nodes or distant organs, giving rise to sec- 
ondary tumors. 

Many of these characters of tumor growth must be referred to mechanical 
crowding of cells, and the demands for nutrition, influenced by whatever 
remnant of tissue organization the tumor may retain. 

The local extension of malignant tumors gives rise to many changes in the 
invaded tissue and reveals many specific qualities of tumor-cells. 

Local extension involves the passage of tumor-cells first into tissue spaces 
and lymph radicles. Whether tissue spaces directly communicate with 
lymph radicles (Reckhnghausen) , or lymph radicles form a series of closed 
anastomosing channels (Ranvier), there is no doubt that inflammatory and 
degenerative processes soon open the lymphatic vessels to the tumor-cells. 
From this point the ameboid properties of the cells (Klebs, Waldeyer), 
mechanical pressure, and the natural course of fluids facilitate the progress 
of the tumor toward the larger lymph-vessels and nodes. Invasion of the 
terminal vessels occurs through the communication between the lymph and 
blood capillary plexuses (Sabine) . 

Through gland ducts, renal tubules, bile ducts, bronchioles, and tubular 
structures also, as well as through vessels, local extension may occur. 

In muscle, extension of tumor-cells is affected by the muscular contrac- 
tions, and especially in pyloric cancer the extension may be widespread before 
any local tumor develops. Blumenthal attributes the invasive properties 
of tumor-cells to their proteolytic enzymes which he conceived as dissolving 
the tissue barriers, but neither the histology of the process nor the known 
capacities of tumor ferments support such a view. 

Inflammatory reaction frequently meets the invasion of tumor-cells. 
This is usually of a low grade, causing the appearance of lymphocytes, large 



mononuclear leukocytes, plasma-cells, and in certain cases polynuclear leu- 
kocytes. It is a highly significant feature of malignant tumor growth and 
must be regarded as a defensive process. Well-marked reaction signifies a 
pronounced capacity to limit the tumor growth but not that the effort will be 
successful. Yet about many epitheliomas which are making little progress 
one finds a thick barrier of lymphocytes. In carcinoma of the breast one 
may find islands of tumor-cells inclosed in masses of lymphocytes and 
presenting clear signs of degeneration. 

It is probable, as Orth concludes, that some degeneration of tumor-cells 
always precedes the tissue reaction. 

Very actively growing malignant tumors and their late metastatic 
deposits often fail to show any inflammatory reaction, but the majority 
of infiltrating growths elicit considerable cellular response from invaded 

Fig. 8. — Diffuse lymphocytic infiltration in a rapidly growing carcinoma of breast. 

tissues. Degeneration and calcification of tumor-cells, formation of macro- 
phages from fibroblasts, endothelium, and leukocytes, and appearance of 
giant tumor-cells, accompany the process. 

Besides the general inflammatory reaction, which fails only in very rapid 
invasion, the different tissue cells exhibit interesting reactions against 

Striated muscle-cells exhibit a relative immunity to many cancers which 
infiltrate the perimysium while long sparing the muscle-fibers. Eventually 
the sarcolemma is penetrated and the tumor-cells pass rapidly along the 
sheaths, abs>Drbing the muscle substance, hollowing out central portions of 
fibers, causing fatty and hyaline degeneration, fragmentation, swelling of 
ends of fibers with formation of giant-cells, and exciting proliferation of 
nuclei of sarcolemma and muscle-cell (Fujinami) . 


Smooth muscle-cells also are rather resistant but are frequently destroyed 
chiefly with simple atrophy. 

Specialized gland epitheUum is usually passive before the attack of 
invading tumors. The ceUs degenerate and atrophy, the membrana propria 
often persisting to inclose pseudo-ah-eoU of tumor-ceUs. Gland ducts are 
often filled or distended with tumor-ceUs which replace the lining cells or 
stimulate them to feeble multiplication. 

Squamous epithelium tends to hj-pertrophy when invaded by alien 
cells, with thickening of the ]\Ialpighian layer, formation of new papillae 
and hyperkeratosis. "With melanoma the swoUen hj'perchromatic and pig- 
mented epithelium is often difficult to distinguish from the tumor-cells. 

Endothehum usually degenerates, exfohates and disappears, or forms 
macrophages. In blood-vessels it may prohferate and assist in the formation 
of stroma. 

In connective tissues where the changes are not obscured by inflammation 
the invasion leads to degeneration, compression, and atrophy of fibroblasts, 
sweUing, edema, fragmentation and solution of fibrils and matrix, and de- 
struction of elastic fibers. 

In manv organs preexisting stroma, vessels and trabeculae, may long 
persist, as in the lung where the air vesicles may be passively filled with 
invading tumor-cells. Strongh- encapsulated organs as testis, kidney, 
lymph-nodes, spleen, may long resist invasion and rupture by tumor-ceUs. 

Reactive growth of invaded connective tissue plays an important 
part in the course of desmoplastic tumors. In scirrhus and metastatic 
carcinoma it may far exceed the bulk of tumor-cells and may eventually 
lead to complete scarring of large quiescent areas once the seat of active 

In fat tissue cells penetrate readily, replace the fat in the tissue cell 
envelops, incite mucous degeneration, form foreign body giant-cells about 
free fat, and somelimes induce proliferation of fat-cells. 

Periosteum and perichondrium ma}- successfully resist invasion from 
without, but when the element of pressure is added they yield to the invading 
cells. From within the attack on bone is more rapid. The matrix is 
absorbed, the tumor-ceUs forming lacunae Uke osteoclasts, dissolving cancel- 
lous tissue, and in many cases of myeloma, sarcoma and metastatic hyper- 
nephroma and carcinoma, perforating the shaft. New blood-vessels accom- 
pany these inroads and new fibroblasts act as osteoclasts, facilitating the 
absorption. Formative osteitis accompanies many tumor processes in bone. 
Located in periosteum it may lead to bony encapsulation on the peripher\' 
keeping pace with bone absorption within. Epithelial tumor-cells invading 
bone are usually free from new bone formation but in a group of carcinomas, 
especially of the prostate, bone-cells are excited to extensive prohferation 
possibl}- e\'en to the extent of a secondary tumor growth. Such cases are 
wholly different from the tumors originating in bone-cells and yielding the 
extensive group of bone sarcomas. In certain very vascular tumors the bone 
is passivel}' absorbed and replaced by large sinuses fined by giant-cells. 

Nerve-trunks owing to their rich lymphatic sheathing are extensively 
invaded, especially by epitheHal tumor-cells. The medullary sheaths 
atrophy from pressure, the axis cjdinders being the last element to disappear; 
while the nuclei ma}- feebl}- multiply. 

In a few instances the nerve-trunks seem to lead the wav in local exten- 
sions especially of epidermoid carcinoma. Brain tissue is quite passive to 
the invasion of tumors, suffers pressure atrophy with degeneration of all the 
specific elements, while necrosis often occurs from occlusion of vessels. 


In the local extension of malignant tumors the inherent momentum of 
cell-growth is doubtless the chief factor. As this property constitutes the 
great distinction between normal regeneration and malignant neoplasia, 
its analysis is of first importance, and will be discussed under general etiology. 
Here it may only be mentioned that the histological study of invaded tissues 
leads to the conclusions above stated, and offers little basis for the theories 
of attraxins drawing the cells deeply into the tissues, or of special toxic agents 
preparing the tissues for invasion, or for the belief that inflammatory changes 
greatly influence the invasive properties of the cells of an estabUshed tumor. 

Functional activity in tumors may be retained in part or exhibited in 
exaggerated form. Geotropism, the quality of living- cells to cover and grow 
along surfaces, is shown in epithelioma, endothelioma, and adenoma. Phago- 
cytosis of bacteria, leukocytes, red cells, fragments of organ cells, other 
tumor-cells, even in mitotic division, bile, fatty material and foreign bodies, 
is exhibited in great variety. Many tumor-cells may exhibit slow ameboid 
properties, as first observed by Virchow in chondroma. 

The abundant matrix in many osleo- and chondrosarcomas, the mucus of 
myoma, fat-cells in liposarcoma, and the rigid alveolar structure of adenoma 
destruens even in metastases, are famihar indications of functional activity. 

Clear examples of overfunction of tumor tissue are seen in mucous 
carcinoma of the colon and in melanoma of the skin. Some tumors of the 
hypophysis seem to be followed by overactivity of the influence of this gland 
on growth. 

While tumors destroying the adrenal, pancreas or thyroid may not be 
followed by Addison's disease, diabetes, or myxedema, yet the functional 
capacities of such tumor tissue have not been demonstrated by chemical 
and physiological tests, and the absence of specific disease with these tumors 
may be referable to the wide hmit of safety of the organs involved or the 
presence of aberrant glands. Hansemann reports complete destruction of 
the pancreas by carcinoma without diabetes although diabetes is common 
with cancer of the pancreas. The early bilateral adenomas of the adrenal 
may very well fail to give Addison's disease, but I have seen complete destruc- 
tion of both adrenals by diffuse hypernephroma with bulky metastases but 
no signs of Addison's disease. The disease was rapid and no search was made 
for aberrant adrenals. 

Tetany followed extirpation of one-half of the thyroid for cancerous goiter 
by V. Eiselsberg, but disappeared on the development of a metastatic focus. 
When this focus was removed the symptoms of tetany recurred, in spite of 
new metastatic growths, these, Hansemann suggests, being more anaplastic. 
Since a large portion of the thyroid was not removed and since the symp- 
toms suggest involvement of the parathyroid rather than the thyroid T 
think the interpretation of this case is very uncertain. Many thyroid 
adenocarcinomas contain colloid and functionate, but it is difficult to find 
any histological traces of functional activity in true cancer of the thyroid. 

Adenoma and primary carcinoma of the fiver usually fail to show traces 
of bile formation, yet Heller observed bile in the cells of a pulmonary metas- 
tasis from fiver cancer. Hansemann found bile in the adenomatous portion 
but none in the carcinomatous areas of an adenocarcinoma of the liver. 

Chemical studies in general support the view that many tumors retain 
some of the function of their parent cells, as shown by the high content and 
quaKties of lipoids in hypernephroma (Wells) and the considerable content 
of iodine in thyroid cancer (Beebe, Ewald). As a rule, functional activities 
diminish with increasing anaplasia, and original overactivity may be suc- 
ceeded by complete failure, as in the pigment-free metastases of melanoma. 


Stroma. — Benign or encapsulated tumors provide themselves with a new 
stroma, while infiltrating growths appropriate and alter the preexisting 
stroma of the invaded tissue. In sarcoma it is usually much less abundant 
than in carcinoma and may consist entirely of blood-vessels, while in cancer 
it is more abundant and fails only in very anaplastic growths. In sarcoma 
a fine intercellular stroma is in some degree constant but may be difficult 
to demonstrate. In carcinoma the cells are in immediate apposition without 
intervening fibrils. Endothelioma reveals its close embryological relation to 
sarcoma by fine fibrils which imperfectly penetrate the cell masses (Woolley). 

In benign organoid tumors the neoplastic process is usually limited to 
one type of cell, in adenoma to epithehal cells, and the new stroma results from 
a formative influence on the cells of the connective tissue and vessels asso- 
ciated with cells originating the tumor. 

Mechanical influences may be prorm'nent in the behavior of this stroma 
or the cells may show active mitosis, and even participate in the tumor growth. 
Thus stroma may have adult characters, or as in fibro-adenoma it may be 
difficult to determine whether the epithehum or the stroma cells dominate 
the neoplastic process. In certain ovarian adenomas the periacinar stroma 
may be extremely cellular, as in sarcoma. 

It seems necessary to sharply distinguish between those tumors whose 
stroma is clearly not neoplastic and those in which it constitutes a, part of 
the tumor and which must therefore be regarded as mixed tumors. In 
chondro- and osteosarcoma the stroma is a combined product, consisting 
of intercellular substance derived from the tumor-cells and ne'W vessels 
derived from the tumor or from peripheral tissues. In teratomatous cancer 
of the testis the stroma may be richly infiltrated by lymphocytes and even 
contain lymph follicles, yielding a type of stroma not seen in any other car- 
cinoma. In veins the stroma of an invading tumor may consist of capillaries 
arising from the vessel-wall. 

Elastic fibers are seldom reproduced in the stroma of tumors and if 
present belong chiefly to the preexisting tissue (Williams). 

According to Bonney they disappear in all areas of connective tissue 
proliferation and cellular exudation, but in isolated areas in breast cancer 
he found bulky periacinar masses. 

The stroma of an original growth may fail to be reproduced in metastases 
with increasing anaplasia, but often reappears in typical form, as in scirrhus 
carcinoma, or in the overdeveloped vessels of fungus hematodes. 

The stroma of surrounding tissues may be (i) passive during the increase 
of tumor-cells, or it may be incited to (2) reactive growth, and there is 
considerable evidence to show that this reaction may reach almost if not 
quite to a (3) neoplastic grade. 

Most highly organized tissues are passive during tumor invasion, as the 
Hver in cancer. Very active growth of tumor-cells usually fails to excite any 
reaction even from simple tissues, as in lymphosarcoma in the skin. 

Apart from collateral hypertrophy as previously discussed, epithelial 
tumor-cells exert a notable formative influence on connective tissue; they 
are desmoplastic. 

So marked is this influence that many carcinomas are composed of much 
new connective tissue and relatively few cells both in primary tumors and 
in metastases. Originally this action is the result of a formative stimulus of 
epithelium upon connective tissue, but eventually it proves protective and 
some cancerous areas undergo complete regression from overgrowth of 
fibrous tissue. 

An osteoplastic influence is exerted on bone tissue by some cancers, 


and active budding and proliferation of muscle-cells is occasionally seen with 
rare myoplastic tumors. 

That the presence of a tumor may excite its stroma or surrounding tissue 
to neoplastic growth has never been demonstrated. While improbable, such 
an action is not impossible and has been suggested by several appearances 
encountered with spontaneous and transplanted tumors. Langhans con- 
sidered the possibility that teratoma testis might incite the spermatoblasts 
to tumor growth in certain complex tumors of this organ. 

In certain lymphocytomas it is dif&cult to distinguish between the 
infiltrative growth, multicentric origin, and the contact influence of tumor- 
cells. Haaland thought he observed the transformation. of pulmonary cells 
into tumor-cells about metastatic carcinoma in the mouse. Ehrlich, Loeb, 
and others have observed the sudden appearance of sarcoma in the course 
of transplantations of carcinoma in mice and rats, and Murray has strongly 
substantiated the belief that this sarcoma arises from the stroma and not 
from the altered cancer-cells. 

The influence of transplantations renders these observations inapplicable 
to human tumors, and several uncertainties still surround the interpretation 
of this remarkable change of structure. No such stroma reaction has been 
observed in spontaneous tumors, although the reactive overgrowth of invaded 
tissues may at times exhibit some of the qualities of a neoplasm. 

Regressive changes occur in tumor stroma very similar to those observed 
in the parenchyma. In some cases these changes greatly influence the 
character and course of the tumor. In progressive keloids the hyaline 
stroma is the chief portion of the tumor. 

Cylindroma or syphonoma, are terms applied to a tumor structure in 
which the stroma appears in the form of elongated twisted thickened cords 
of hyaline material. This structure is most often seen in basal cell carcino- 
mas. In old regressing carcinoma large areas of tissue may show very few 
tumor-cells in a mass of quiescent hyaline stroma. 

Edema, mucous degeneration, and calcification of stroma are frequent 
factors influencing the course of tumors. 

In cancers, fatty degeneration often affects not only the parenchyma but 
also the stroma. Calcification may begin in the stroma and gradually 
involve the entire substance of myomas, or it may form a hard shell about 
this or other growths. Solid or lamellated concretions of hyaline material 
of uncertain nature are occasionally seen in the stroma as in the parenchyma. 

Experimental Analysis of Stroma Reactions. — Much light has been thrown 
on the meaning of the cellular reaction to invading tumors by experimental 
studies on transplantable tumors. 

The inoculation of susceptible animals with a tumor graft is usually followed by a 
brief local polynuclear leukocytosis incited by diffusible products of tumor-cells, red cell 
detritus, and traumatism. 

An excessive degree of this reaction may prevent the graft from surviving. With 
successful grafts there is later a gathering of lymphocytes in moderate number and these 
may form rich foci which interfere with the growth of the tumor-cells. The graft then 
becomes successfully implanted by inciting the growth of fibroblasts and capillaries about 
which the cells adhere. Often the vessels of the graft are appropriated and reopened. 
From this point the growth proceeds by multiplication of the transferred cells about the 
new stroma derived from the host's tissues. In its further progress the tumor may become 
encapsulated and thus be shielded from cellular attack, or it may early show infiltrative 
tendencies unopposed by any cellular reaction. 

In some conditions animals become immunized by the temporary growth foUowedby 
spontaneous regression of the tiimor (Gaylord, Clowes). In such cases one of the earliest 
changes observed is a focal degeneration of the tumor-cells accompanied by an accumula- 
tion, probably from chemotactic influences, of ameboid large and small lymphocytes. 


mononuclear leukocytes, and often many plasma cells. Minute hemorrhages may occur in 
the tumor about which eosinophile cells gather. These cells surround the tumor and 
penetrate within it, sometimes richly infiltrating the stroma. 

Then follows contraction of the tumor mass, multiplication of spindle-ceUs, growth of 
new vessels and encapsulation of the regressing tumor. During the absorption process 
the endothelial macrophages appear and giant-cells result from fusion of tumor elements. 
Even to a late stage mitosis may persist in such encapsulated regressing tumors and 
section of the capsule may excite renewed growth. 

In immune animals the implantation of cancer excites a more active emigration of 
polynuclear cells, the degeneration of parenchyma cells appears earlier and is more pro- 
nounced, and the lymphocytic reaction is less marked, plasma-cells being absent (De Fano) . 

With a sarcoma of the hare transplanted into rabbits v. Dungern and Coca obtained 
in the rabbit hypersensitization followed by complete immunity by successive implantations 
of the tumor. The second or third graft excited a rich production of endothelial macro- 
phages which promptly surrounded and englobed the degenerating tumor-cells, and formed 
macrophage thrombi in the small vessels. Leukocytes were not prominent in this reaction. 
In the adjoining lymph-nodes were many similar endothelial macrophages. 

With Jensen's mouse cancer Russell found that grafts in resistant mice suffered ex- 
tensive degeneration and the peripheral rim of surviving cells failed to excite a stroma 

Working with Japanese mice Burgess observed that in both susceptible and non- 
susceptible animals tumor grafts established themselves by exciting a stroma reaction, 
after which in the immunized mice the nutrition of the tumor was cut off by overgrowth of 
fibroblasts with much intercellular substance. 

In the interpretation of these results one may conclude that in highly immunized and 
hypersensitized animals the reaction is pronounced, immediate, and exudative, preventing 
the growth of new stroma or by thrombosis (v. Dungern-Coca) shutting off blood-supply. 
In spontaneously immune or in alien animals the reaction is less pronounced, new stroma 
forms, but later degeneration of the graft leads to strangulation by connective tissue 

Analyzed in the light of experimental studies, the processes connected 
with the growth of tumors signify as follows : 

Infiltrative growth without stroma or cellular reaction indicates absence 
of protective forces in the body. 

The development of vascular stroma from the host's tissues, up to a 
certain degree, is favorable to the growth of the tumor. 

Overgrowth of stroma with much intercellular substances is a menace to 
the tumor. 

Polynuclear leukocytes gather about the tumor as the result of necrosis, 
trauma, bacterial processes and actively chemotactic tumor products, and 
signifies a sharp response against the tumor. 

Lymphocytes and large mononuclear leukocytes signify a reaction of 
immunity, limitina; the growth of cells, estabhshing local tissue immunity, 
and conveying ini i aune forces to a distance. Plasma-cells are a later factor 
in the same process. 

Eosinophile cells gather about hemorrhages. 

Endothelial proliferation, if slight, figures as a part of the stroma reaction, 
but if excessive, macrophages form and oppose cell-growth. 

Degeneration of tumor areas, fusion giant-cells, and necrosis, may result 
chiefly from withdrawal of nutriment. 

It is obvious that most of these conclusions have been warranted by 
previous study of human material, but it must be claimed that the experi- 
mental studies have rendered them more precise. ' i 

Blood-vessels. — The relation of blood-vessels to tumor growth presents 
two highly important aspects, first in connection with the beginning of 
cancer, and second in the channels of nutrition in the progressive growth of 

Franz Boll from a study of early epithelioma and of the growing edges 
of such tumors concluded that changes in the blood-vessels determined 


the proliferation of epithelium. He found the walls of capillaries thickened, 
opaque and varicose. The adventitia of veins and arteries was invaded 
by cells, and the vessels lost their sharp demarcation from the surrounding 
tissue which took on an embryonal character. 

These changes were found throughout the tumor and preceded the exten- 
sion of the growth. An extensive disease of the vascular system of the whole 
area involved was in Boll's opinion an essential part of the cancerous process. 
The later study of the vascular system of mahgnant tumors, while not fully 
demonstrating the prime influence of vascular lesions, has yet emphasized 
the importance of the blood-supply in the origin and growth of many 

TheiDiaber, among recent writers, has emphasized the importance of 
arteriosclerosis in the origin of many cancers in elderly persons. He con- 
cludes that the local predisposition of aging tissues to cancer results from a 
diminution of the cellular elements and a contraction of the blood-vessels 
in the connective tissue. On the other hand Warner reports a study of the 
blood-vessels of 206 cancerous organs, finding that 105 (51 per cent.) showed 
obstructive arterial changes, 118 (57 per cent.) fibrosis, and 85 (42 per cent.) 
lymphocytic infiltration. Since normal vessels were present in almost half 
the cases endarteritis could not be considered a necessary factor in the 
production of the disease. 

Possibly a very close analysis by various methods of the circulatory 
conditions about beginning carcinomas would show a more constant relation 
between this factor and carcinoma in elderly persons. On the contrary side, 
increased vascularity has been demonstrated in the beginnings of many 
carcinomas, and has been assumed to account for the awakening of many 
cell rests. Evidently the exact relation between blood-supply and the 
inception of carcinoma has not yet been elucidated. 

Since the demands for nutrition in tumors exceed those of normal tissues 
the significance of the vascular system is correspondingly increased. In the 
older literature S. von der Kolk stated, as a result of injection experiments, 
that only arteries formed anew in tumors, the veins being those which became 
inclosed in the tumor. 

Extensive lesions of the blood-vessels have been described by many 
authors as a significant feature of tumor growth (Eberth, Belapolsky). 
Virchow pointed out that the circulation in many tumors is to some degree 
independent of and different from the general circulation, a fact according 
with the parasitic nature of tumor growth. The difficulty of passing injec- 
tion material from arteries through veins he attributed to compression, 
invasion, and thrombosis of the veins and to an indirect connection of veins 
with arteries. 

The "saber-sheath" flat veins oii the surface of many tumors result 
chiefly from compression. The turgescence of enlarged veins leading out 
from malignant tumors was one of the gross features giving rise to the term 
cancer. Compared with the affected organ the vessels of actively growing 
tumors are more numerous, dichotomous branching of the arterial system is 
absent or irregular, and the main vessels are often tortuous and spirally 

The vascular system may be studied by corrosion preparations, by i--ray photographs 
of specimens injected with bismuth and oil, and by Hill's method of injection with India 
ink with subsequent clearing of tissues by alkali and glycerine. 

From results obtained by the two latter methods Goldman observed the marked increase 
in vascularity in the periphery of growing tumors, while in quiescent cases the central 
portions were nearly devoid of vessels and those of the periphery reduced. The new vessels 


were usually of small size, tortuous, irregularly distributed, and the arteries rapidly broke 
up into capillaries. 

In the growth of blood-vessels in tumors one may distinguish two factors, 
direct irritation by tumor products and a formative influence exerted by 
tumor-cells. Along the growing edges of aldvancing carcinoma there is often 
a zone of h5^eremia and section shows many new capillaries and round- 
cells, producing at times a structure which is equivalent to granulation 
tissue. Such changes, which soon regress, may be referred to inflammatory 
reaction. In the artificial implantation of tumor -grafts, in the organization 
of intravascular tumor thrombi, in the natural growth of angiomas, and in 
the recurrence of many varieties of cellular tumors, new vessels develop 
apparently as a result of a formative influence centered on the vessels of 
surrounding tissues. 

In many endotheliomas and some angiomas the vessels constitute the 
unit of the tumor; elsewhere they are devoid of neoplastic qualities. The 
structure known as perithelioma may develop solely as an expression of a 
formative influence of tumor-cells upon blood-vessels, and has been observed 
in angiosarcoma, fibroblastic sarcoma, h){pernephroma, epithelioma, and 

The structure of the blood-vessels varies greatly. 

In some benign growths the large arteries exhibit all three coats in pro- 
portions approaching the normal. In most malignant tumors muscle- 
tissue is deficient in walls of arteries, and the main type of vessel is composed 
of adult fibroblastic tissue lined by endothehum. In many endotheliomas 
the vessels are chiefly spaces lined by tumor-cells. In adenoma of thyroid 
and some primary carcinomas of the liver widely dilated capillaries lined by 
endothelium produce highly vascular growths. In telangiectatic sarcoma 
the capacious sinuses may be lined by endothelium backed by connective 
tissue. In bone-sarcoma widely dilated sinuses are lined by tumor giant- 
cells, as in the pulsating sarcomas of tibia and humerus. Finally a feeble 
circulation may be maintained in intercellular tissue spaces without definite 
lining cells. 

Lesions of blood-vessels are frequent and important for both the tumor 
and the organism. Functional inadequacy or compression of vessels is 
chiefly responsible for the degeneration and necrosis of the parenchyma of 

Rupture from degeneration or trauma gives rise to hemorrhage and 
infiltration with blood. Strangulation of pedunculated ovarian cystomas 
or uterine myomas, cutaneous sarcomas, and even breast cancer may lead 
to complete hemorrhagic infarction, necrosis, and sloughing. 

Hyaline degeneration of vessfl-walls yields a characteristic structure in 
certain sarcomas, as in the ovary. 

The vessels of invaded tissues are passively compressed or react with 
periarteritis and round-cell infiltration or the walls are invaded by tumor- 
cells by way of the vasa vasorum, and the perivascular lymphatics form a 
channel directing the course of invading cells. Arterial trunks long resist 
the attack of most tumors, but fall a ready prey to squamous epithelioma 
whose cells penetrate the adventitia, split up the muscularis, occlude the 
lumen, and thus cause necrosis of the invaded tissues. In such vessels the 
tumor-cells may largely disappear with fibrous organization. In the veins 
tumor-cells readily penetrate to the intima producing nodular swellings, 
endophlebitis carcinomatosa, and eventually rupturing into the vein. 

It has often been noted that very vascular tumors may long fail to develop 


metastases, yet the invasion of tissue veins is the chief source of metastatic 
growths in sarcoma and with many carcinomas. An intravascular position 
characterizes the growth of certain tumors throughout most of their history. 
Chorioma of the uterus and testis is remarkable for its limitation to pelvic, 
vaginal, abdominal, and pulmonary veins. According to Marchand and 
Risel other rare tumors growing in or invading the blood-vessels may imitate 
the hydatidiform appearance of chorioma. Hypernephroma often passes 
up the renal vein, and distends the vena cava and the hepatic veins before 
generalizing. An adenocarcinoma of the liver regularly appears as a circum- 
scribed growth distending the large branches of the portal vein in this organ. 
Course and Rate of Growth. — Many variations are observed in the course 
and rate of growth of tumors. 

1. Rapid and progressive multiplication of tumor-cells without notable 
intermission or hindrance is characteristic of most highly malignant 

Lymphosarcoma in adult or young subjects may run its course in a few 
weeks with extensive metastases in many organs. Cancer of the liver in 
elderly subjects may terminate with steadily progressive symptoms in three 
months. Some melanotic tumors after becoming established run a very 
rapid course. Children are especially susceptible to such forms of fulminant 

Pregnancy usually accelerates tumor growth. I have observed cancer 
of the breast arising after gestation prove fatal, with widespread metastases, 
before term. With experimental tumors pregnancy sometimes accelerates 
but at others retards growth (Moreau, Loeb). 

2. Intermittent periods of relative or complete quiescence occur in the 
course of many malignant tumors. It is probable that such quiescence is 
often more apparent than genuine. The progress of the edges of carcinomas 
and the involvement of lymph-nodes often appear to be sudden and the 
growth reaches certain dimensions in a few weeks after which progress is 
much slower. Exhaustion of growth capacities may partly account for 
such cases, but usually it is referable to pressure on a distended capsule or 
hindrance to the increased blood-supply. The onset of edema, hemorrhage, 
retention of secretions, inflammatory complications, or central necrosis, 
fibrosis, contraction, and evacuation of fluids, simulate changes in the rate. 
of growth. Or a primary tumor may be at a standstill while internal meta- 
stases are active. 

Yet apart from these conditions there appears to be a frequent tendency 
in certain tumors to grow steadily to a certain bulk and then to remain sta- 
tionary until trauma or other external factors initiate a new period of growth. 
A definite form of intermittent growth is observed in cancer of the stomach 
and epithelioma of mucous membranes which remain localized ulcers for a 
time, but suddenly extend more rapidly upon the invasion of liver or lymph- 
nodes. Malignant change in a benign tumor is marked by sudden accel- 
eration of growth, as when myosarcoma develops on a uterine myoma. 

3. A natural limit exists to the course of some tumors and the rule of 
endless increase is by no means invariable. Myomas of uterus may cease 
to grow at the chmacteric. Warts and Upomas may be self-hmiting. The 
relative bulk is in part an indication of the general momentum of growth of 
tumors. Thus squamous epithelioma does not as a rule produce a bulky 
growth, and seldom yields bulky metastases. Epithehoma of skin or even 
of esophagus may produce ulcers which after steadily growing to certain 
dimensions long remain practically quiescent, and on section one finds a 
firm basis of connective tissue beneath the ulcer and few or no mitoses in 


the tumor-cells. Martland has reported a mixed tumor of a salivary gland 
of 20 years' duration and long stationary. The same numerous and com- 
plex factors which interrupt growth may succeed in estabUshing a perma- 
nent limit. It is also beginning to appear that a local or general immunity 
may be slowly established and effectually prevent growth. 

4. Spontaneous cure is observed with certain tumors under peculiar 

There is a considerable series of polj^oid tumors of skin and mucous 
membranes where nutrition is cut off by constriction of the pedicle and which 
are either completely extruded or regress without detachment. Fibrornas 
and chondromas may accidentally compress their nutrient vessels with 
regression or sloughing.- 

Within the tumor hyaline degeneration of vessels and mucous degenera- 
tion of stroma and parenchyma usually retard growth and may completely 
exhaust the capacity for growth. Calcification of parenchyma and capsule 
may incarcerate the tumor. 

Inflammation about the borders and in the body of the tumor seems to 
cause the occasional extrusion of certain sarcoid growths of the skin (Ran- 
dolph, Watson). Accidental infection by erysipelas has been followed by 
spontaneous regression of sarcomas. 

For somewhat uncertain reasons benign tumors have been known to 
regress, as exostoses (Nasse, Starck) and papilloma of the bladder (Nitze). 
A local or general immunity seems to be the only explanation of an increasing 
list of spontaneous cures of highly malignant tumors. 

Most of these cures are only partial. Among them are sarcomas, squa- 
mous epitheliomas, and glandular carcinomas (Czerny) . 

Tripier records a case of multiple round- and spindle-cell sarcomas of 
wide distribution with pulmonary hemorrhages which disappeared after the 
condition seemed hopeless. The exact nature of this case is uncertain. 
Kaposi has recorded the disappearance of lymphosarcoma of the upper jaw 
while internal metastases were in progress, and Reichel, of a spindle-cell 
sarcoma of forehead. 

Some of the cures have followed persistent recurrence after operation. 
(Shepherd, Watson). In dermatological literature there is a considerable 
list of reported spontaneous cures of cutaneous sarcomas. While the true 
nature of many of these processes is uncertain, I do not believe that they 
can properly be rejected, as Williams would, from the field of neoplasms. 

Spontaneous recovery from chorioma belongs in a special category owing 
to the peculiar relations of this tumor. In 15 cases this process has been 
known to regress after the occurrence of pulmonary or vaginal metastases, 
in three of which neither uterus or vaginal nodules were removed. In 7 
other cases recovery followed partial removal (Ewing) . 

Recovery after partial removal of ovarian adenocarcinoma is not infre- 
quent. Rotter has described a case of complete cure of an original malignant 
adenoma of the rectum, although after 3 years a metastasis of the iliac 
bone was found at autopsy. Here a complex form of local tissue immunity 
seems to have existed. 

With cancer of the breast there is a group of cases in which the disease 
becomes locally extensive and internal metastases form, the patient becoming 
greatly debilitated, but at this stage the picture changes and very marked 
improvement sets in attended with regression or even disappearance of 
many tumors. Although these patients usually die of the disease they may 
enjoy good health for many years. General debility cannot account for such 
a course, and there seems to be no other explanation than the development 


of a relative general immunity. In MacKay's case improvement was coinci- 
dent with absorption of a pleural exudate, and I have followed several very 
resistant cases which were marked by continuous ascites, sometimes chyli- 
form (Hodenpyl). 

Somewhat similar are those cases in which primary tumors disappear 
and the patient long enjoys fair health, during the very slow progress of 
internal metastases (Bryant, Gould, Broca, Hutchinson). 

Fig. 9. — Almost complete replacement of liver by metastatic fibrosing carcinoma of breast. 
The external tumors in this patient had notably regressed. 

The self-limiting tendencies of certain epitheliomas seem to have resulted 
in spontaneous cure in cases reported by Sanger. Many interesting cases 
of this general type are collected by Williams. 

The morphology of spontaneous regression presents special characters 
with each t)rpe of tumor, among which are the calcification and giant-cell 
formation in epithelioma (Petersen), giant-cell degeneration in chorioma 
(Teacher) (Ewing), extreme hyalinosis in cancer of the liver, and simple 
atrophy of parenchyma in the large cell tumor of the testis. 


The Significance of Malignancy. — The distinctions between benign and 
malignant tumors involve questions of great interest from both the theoret- 
ical and the practical sides. If malignancy were a purely clinical conception 
it would be impossible to draw any rigid distinctions between benign and 
malignant tumors since nearly all tumors may occasionally prove fatal. Yet 
the tendency is to restrict the term to tumors which exhibit certain features 
which are essentially deleterious to the host. The most important of these 
features are infiltrative growth, local destructive properties, recurrence after 
removal, formation of metastases, local interference with function and 
general toxic action of absorbed tumor products. These elements involve 
both anatomical and clinical effects. 

Infiltrative growth is the most important of the anatomical factors in 
malignancy. This property facilitates local arid general extension, renders 
removal difficult, is responsible for many recurrences, and is a constant 
preliminary to the local destruction of tissue. The -controlling influence 
of encapsulation is seen in the harmless course of many very cellular tumors 
which may readily be enucleated while circumscribed but which on rupture 
of the capsule take on infiltrative and malignant properties. It is also 
demonstrated in transplantable tumors which upon incision of the capsule 
may be transformed from regressing into malignant infiltrative growths 

During experimental increase of growth Bashford and others have shown 
that originally circumscribed mouse tumors take on infiltrative qualities. 
Yet Borst notes that early fibroma of the kidney showing infiltrative 
growth later becomes encapsulated and the same observation applies to 

Rapidity of growth is usually associated with infiltrative qualities and 
deleterious effects and if carefully separated from simple increase in size 
from other causes is a nearly constant sign of a dangerous tumor. Local 
destruction of tissue is a variable feature prominent only in certain types of 
tumors, appearing early in squamous epithelioma, and constituting the sole 
malignant quality of rodent ulcer. In benign tumors destruction of tissue 
is a secondary effect of pressure. Hemorrhage from destruction of vessels 
may be the chief instrument in the cachexia which marks the course of maiiy 
ulcerating tumors as in the stomach and uterus. Pain from involvement 
of nerves is often the first and even the last clinical problem in the course 
of a lethal neoplasm. 

Formation of metastases may be held to constitute any tumor malignant. 
Being usually the result of infiltrative growth, invasion of vessels, and cel- 
lular character, and signifying dissemination and collapse of defeilsive 
powers, "it is the most impressive external sign of malignancy. It is, febwever, 
by no means constant in lethal tumor processes, as in gUoma, rodent ulcer, 
etc., and certain tumors otherwise benign may occasionally yield distant 
secondary growths, as chondroma and myoma. 

Local interference with function is the chief dangerous feature of many 



gliomas, epitheliomas of larynx, esophagus, cardia, and cancer of the pylorus 
and rectum, and many benign tumors become serious chiefly from effects 
of pressure. 

The general intoxication resulting from tumor growth is a complex subject 
which has long been recognized as one of the most obscure and important 
problems in the natural history of malignant tumors. Hemorrhage, mechan- 
ical interference with nutrition, pain, the psychical condition, abnormal 
secretions, destruction of important tissues, the toxic action of products 
absorbed from degenerating and ulcerating areas, and bacterial infection 
combine to produce the cachexia of tumors and require consideration in a 
separate chapter. When the clinical signs of malignancy declare themselves 
the conditions exciting them often belong to the past history of the disease 
and are often irremediable. 

Diagnosis of Malignancy. — The prediction of the course that a given 
tumor will take is based upon two sources of information, anatomical and 
microscopical diagnosis and accumulated experience regarding the usual 
behavior of tumors of known histological structure. Fortunately a parallel 
exists to a very marked degree between the histological structure and the 
usual clinical course. 

The main subdivisions of tumors into benign and malignant is accom- 
phshed at once and as a rule only by microscopical structure. Under some 
circumstances the microscopical structure may stand alone and override 
all other considerations. Far more usual is it to find the interpretation of 
structure to be greatly influenced by clinical information regarding the exact 
location of the tumor, its attachments, the presence of a capsule, and the 
age and condition of the patient. 

Pathologists and surgeons alike commonly ignore these essential condi- 
tions of accurate diagnosis, a delinquency which is responsible for the wide 
and firm differences of opinion regarding the relation of microscopical struc- 
ture and prognosis. It rtiust also be urged that tumor diagnosis is commonly 
assigned to the relatively inexperienced, whereas the correct interpretation 
of structure in the light of clinical data requires a very wide experience not 
only with the general tendencies of specific structures but with the observed 
cHnical courses of different tumors. 

A somewhat general estimate of the mahgnancy of tumors may be based 
on the distinctions between adult, and embryonal or anaplastic growth. It 
has long been recognized that the greater the variation in type between a 
tumor and its originating tissue the more malignant the tumor, and Hanse- 
mann has effectively emphasized this principle pointing out that the morpho- 
logical evidences of anaplasia have a physiological significance indicating the 
degree to which the process is freed from growth restraints and the control 
of the organization. Yet here is encountered the difficulty of distinguishing 
between original embryonal quaUties and signs of acquired anaplasia. One 
group of tumors arises from embryonal cells which have lagged behind in 
development and such tumors bear an embryonal stamp. The histological 
signs of this embryonal character are often difficult to distinguish from the 
signs of anaplasia, and if they are wrongly interpreted an erroneous impression 
may be drawn of the mahgnancy of the tumor. The great theoretical value 
of the distinction between embryonal and anaplastic cells is not, however, 
equalled in practical importance, since both types of tumors, especially the 
latter, are usually quite malignant. The histological signs of anaplasia are 
a cellular character, marked variations in size in either direction from the 
originating cells, increase of chromatic nuclear substance, abundance and 
abnormality of mitoses, and loss of polarity and diffuse infiltrative growth 


of cells. In many instances lack of reaction of the tissues against the 
infiltration of tumor-cells is a significant feature. Equally important are 
the general signs of exalted nutrition and vitality of the cells. Upon these 
features one may safely base the estimate of growth capacity and potential 
malignancy of tumors. 

The clinical course does not always accord with the signs of growth ca- 
pacity. Many factors besides the surgeon's knife influence the progress of 
growth. One must distinguish between the potential malignancy and the 
clinical course of tumors. The histological signs of malignancy measure 
the potential malignancy of a tumor but the clinical course is subject to 
wide variations from the position of the tumor, hemorrhage, trauma, changes 
in rate of growth, bacterial infection, etc., any one of which influences may 
greatly alter the course. With these important Hmitations it may be asserted 
that there is a close parallel between histological structure and the malig- 
nancy of a tumor. 

In practice the most frequent source of discrepancy between histological 
signs of malignancy and clinical course arises from the confusion of embry- 
onal with anaplastic qualities. Some of the most malignant looking tumors, 
as the congenital sarcoma of the kidney, teratoma testis, sarcoma of skin, 
adenoma of thyroid, gastro-intestinal tract and other organs, and large cell 
epithelioma of the skin, long remain localized and spare the lymph-nodes. 
These tumors are of embryonal origin, and the growth capacities of their 
abundant atypical undifferentiated cells are much below those of other tumors 
whose cells show true anaplasia and increasing loss of growth restraints. 
Recurrent adult and highly malignant squamous epithelioma yields cells 
with little trace of pavement character, which closely resemble those of a 
benign embryonal epithelioma. Here and in many other fields it is neces- 
sary to know the history and to divine the origin of the tumor in order to 
construct the prognosis. 

Different standards apply to different tissues in estimating the significance 
of cell-growth. The thyroid, liver, uterine mucosa, and lymph-nodes re- 
spond readily to various stimuli and the gross and microscopical criteria 
applicable to those organs cannot be employed for the breast, kidney, or 

The nature and origin of many round-cell growths are obscure and 
difficult to recognize, while their natural history is imperfectly known, so 
that in this group there are many apparent discrepancies between structure 
and prognosis. Finally, there is a natural limit to the onward growth of 
many malignant tumors and it should occasion no surprise if in advanced 
stages certain malignant tumors should advance more slowly, become 
stationary or even regress. Further knowledge may possibly show that 
under some conditions the presence of tumor tissue is a safeguard rather 
than a menace to the organism. 

The transformation of a benign into a malignant tumor occurs in rare 
instances. To be separated from this category are those cases in which 
one element of teratoma gains the upper hand and eventually outstrips the 
relatively benign portions. Apparently this event may occur at any stage 
of the growth of teratomas. In a case which I have reported a teratoma of 
the testicle containing adult thyroid, epidermoid cysts and diffuse carcinoma, 
the malignant process was delayed for years. In an ovarian dermoid which 
had broken into the colon, I have seen epithelioma developing from the irri- 
tated epithelial surface. In another group of cases, a pigmented mole or 
adrenal rest in the kidney after growing slowly for years may suddenly, 


perhaps after trauma, take on malignant qualities. Here the neoplastic 
nature of the originating cell group may be doubted. 

When an encapsulated nearly stationary but potentially malignant 
tumor breaks through its capsule and grows rapidly there is no ground for 
assuming a change in the quality of the tumor. Different portions of many 
tumors differ widely in structure, and many apparent changes in type of 
tumors rest on the mistaken conclusions of a biopsy. 

The impression that the tissue of a well-established benign tumor may 
after a long period take on diffuse malignant characters is based chiefly 
upon observations of uterine myomas. About 3 per cent, of such tumors are 
said to become sarcomatous, but the exact significance of this change has 
always been subject to doubt and discussion. In the great majority of such 
cases the impression that a malignant change has occurred is based on the 
discovery of a myosarcoma by operation on a tumor of which the previous 
history has been imperfectly traced. In a smaller proportion a very slowly 
growing fibroid has been known to exist for years when it suddenly begins 
to grow rapidly and on removal one portion of the tumor is found sarcoma- 
tous. This portion is often connected with an ulcerated surface of the endo- 
metrium, an observation which justifies the conclusion that the ulceration 
and the sarcomatous change are in some way related. In other cases the 
myoma has been cystic or angiomatous. Sarcomatous change of this sort 
has occurred in several authentic cases (Piquand). The true interpreta- 
tion of this change remains uncertain. It may be that the sarcoma results 
from a new tumor process arising in the stroma; or a sarcomatous process 
becomes established in the stroma of the endometrium subsequently in- 
vading the myoma (W. Williams); or the more rapid growth and the cell 
changes represent the natural course of the original tumor which has not 
.thereby acquired any new powers of growth (Borst). The enumeration of 
these possibilities suffices to reveal some of the complications of this subject. 

Carcinoma has been known to develop from the epithelial elements 
present in many uterine fibroids, as in cases recorded by Klob, Roily, and 
others. More often uterine fibroids are invaded by malignant tumors of 
adjacent or distant organs, as by cancer of the endometrium, lung (Schafer), 
or breast (Bender). 

In the breast, cancer and sarcoma have been found in connection with 
fibro-adenoma, and it appears that the malignant tumor has arisen from the 
tissues of the benign. In the case of ovarian cystomas there is no evidence 
to show that varying grades of malignancy exhibited by these usually benign 
tvunors constitute a definite change in type. 

While therefore the transformation of certain benign tumors into malig- 
nant forms has been shown to exist, this event is among the rare occurrences 
in the natural history of tumors. In the fields where it is supposed to be most 
frequent, R. Williams has shown that benign tumors are much less hable 
to change of type than are the normal tissues to develop cancer. This 
conclusion accords with the rule that a tumor process becomes established at 
a certain momentum which it tends to maintain throughout its entire course. 


The emaciation accompanying tumor growth has long been recognized as 
one of the most significant and obscure of its many features. It may be said 
to be a constant effect of both very maUgnant and relatively benign tumors. 
It occurs early or late, with or without anemia though probably always with 
diminution in the total volume of blood, sometimes with preservation of the 


fat deposits, and is preceded by distinct muscular weakness. It affects 
chiefly the muscular system but also all cellular organs and tissues. Inani- 
tion is undoubtedly the chief factor in the loss of weight. Mental depression 
leading to distaste of food, the lowering of digestive capacity, and mechanical 
obstruction to the ahmentary passages, reduce the amount of food absorbed. 
In six cases of uterine cancer v. Noorden calculated the voluntary ingesta at 
300-1200 calories. The importance of starvation is strikingly emphasized 
by the rapid increase in weight following rehef of esophageal and pyloric 

Abnormal loss of proteins figures in many cases, by hemorrhage, exudation, 
albuminuria, and diarrhea. A febrile destruction of proteins occurs very 
often and is a somewhat neglected factor especially in low forms of infection, 
and the difficulty of securing bacteria-free tumor tissue suggests that bacterial 
poisons may often be at work when not suspected. The influence of ac- 
tively regenerating tissues upon the general organism has been generally 
overlooked in discussions of the tissue atrophy of cancer. The exalted ca- 
pacity of tumor-cells to absorb nutriment may long be successfully met by 
physiological processes but there is a point, as mentioned by Cramer, when 
the general body tissues begin to yield their food supply to the tumor. That 
this process may actually lead to atrophy of normal tissues has been shown 
experimentally by Stockard who observed marked diminution of the re- 
maining arms of Medusa during the normal regeneration of amputated 
arms. The restored individuals were sometimes only half as large as the 

The conception of a toxic destruction of protein tissues emanated from the 
observations of F. Miiller, 1889, upon seven febrile cancer patients. In two 
of these, pancreatic, and mammary, a toxic destruction was not apparent 
but in four gastric cases the nitrogen loss was excessive and in a case of. 
cancer of penis with metastases, an intake of 21 gm. N., 3064 calories, failed 
to balance the loss. These observations support the general belief that the 
products of degenerating necrosing and infected tumors cause toxic destruc- 
tion of tissues in cancer as in other diseases. 

Yet the belief in a specific toxic destruction of protein tissues in cancer 
has never been fully accepted. It is worth noting that Wilks in 1868 from 
the study of 2000 autopsies at Guy's hospital concluded that cancer as such 
does not cause cachexia. This conclusion has steadily gained support. 

Many observers have failed to detect the toxic nitrogen loss described by 
Miiller and verified by Klemperer. Many cases with entirely normal protein 
metabolism have been observed by Widal, Setti, Braunstein, Clowes, and 
Lewin, and even a nitrogen retention has been found in cases studied by 
Schopp and Moraczewski. It is possible that a correct interpretation of 
these cases must consider such factors as masking of a nitrogen loss by inade- 
quate excretion by the kidneys, or by the sparing action of carbohydrate diet. 
Schmidt points out many resemblances between the supposed toxic destruc- 
tion of cancer and that of fever, as the sparing of fat, and the high oxygen 
consumption, which is yet not in proportion to the protein catabolism. Yet 
these objections merely emphasize the difficulty of securing a suitable subject 
for investigation, among which gastric, hepatic, intestinal, and ulcerating or 
infected cases must not be included. In a case of mammary cancer which 
seemed to meet all obvious requirements Benedict and Gorslin at New York 
Hospital readily maintained nitrogen equilibrium and found no abnormality 
in the partition except marked increase of creatin over a terminal period of 
10 days. Our present knowledge derived from careful clinical study, 
from observation at autopsy, and finally from chemical investigation, seems 


to warrant the conclusion that a pecuhar toxin secreted by cancer-cells 
and leading to cachexia does not exist. 

This conclusion does not seem to need revision in the light of various 
chemical analyses of cancer tissue (Petry, Wolf, Beebe, Schaffer), since 
no new and abnormal toxic substances have been shown to exist in tumor 
tissue, although quantitative variations in their split products may occur. 

Tumor Ferments. — It is everywhere agreed that cellular tumors are very 
prone to degenerate and that their products may cause intoxication. This 
subject appears in a new phase in the study of tumor ferments. Petry 
emphasized the well-known rapid autolysis of tumor tissue which he referred 
to autolytic ferments. Buxton found quantitative but no qualitative dif- 
ferences between tumor and normal tissue ferments. 

A heterolytic action of tumor ferments has been actively claimed by Neu- 
berg, Jacoby, and Blumenthal and Wolff, but I think on insufficient grounds. 
None of these authors has considered the influence of bacterial or leukocytic 
ferments, factors which seem to me to influence their results and to explain 
their inconstancy, and warn against the hasty acceptance of the compre- 
hensive deductions drawn from them. Kepinow has pointed out the obvious 
defects of this work and with Muller has shown that when free from leukocytes 
and bacteria cancer tissues do not accelerate autolysis in other organs. Petry 
also was unable to find evidence of specific ferment activities in the aseptic 
blood of cancer. Abderhalden and his associates have endeavored to show 
that the splitting of optically active polypeptids in the presence of tumor 
juice follows a different course than with extracts of normal tissue, but their 
results were inconstant. 

The study of experimental tumors in lower animals has given additional 
support to the belief in the innocency of well-nourished tumor tissues. 
Especially in mice it is observed that tumors nearly as large as the animal 
itself may be well borne until the moment when ulceration, infection or 
necrosis occur. 

The Blood in Cancer. — Changes in the blood are of interest to the student 
of tumors chieflyin their relation to cachexia and to the problems of immunity. 

Benign tumors exert little or no influence upon the blood and the very 
slight change of the early stages of malignant tumors may be cited in support 
of the view that in the beginning cancer is a local disease. Likewise in the 
established stages of many malignant tumors the blood remains practically 
normal in quality, and even in advanced stages there are many records 
showing normal cells and hemoglobin. Cabot collected 19 cases of 
gastric cancer with over five million red cells. Laache noted the same immu- 
nity of the blood against the effects of a bleeding uterine carcinoma and con- 
cluded that there is an individual insusceptibility to the effects of malig- 
nant tumors. In some gastric cancers there may be a concentration of the 
blood, probably from diminished fluid ingesta which maintains the quality. 
Leichtenstern, Patrigeon, and the writer have observed cases of gastric 
cancer with increase of Hb. to 100 per cent., shortly before death. Even 
when marked emaciation occurs the blood may fail to deteriorate in quality, 
so that Hampeln, Neubert and others speak of a marantic as opposed to the 
usual anemic type of malignant disease. 

Yet the great majority of malignant tumors are associated throughout 
most of their course with progressive deterioration in the quality and quan- 
tity of the blood. Usually this anemia takes the form of a secondary chlo- 
rotic process, with loss of Hb exceeding the reduction of cells, low Hb index, 
and slight leukocytosis. In a notable group of cancers of the stomach the 
anemia dominates the clinical picture, and takes a secondary pernicious form. 


The sources of the anemia of malignant tumors are numerous and complex 
and being variously combined must be estimated separately for each case. 

Hemorrhage is the most obvious source of deterioration of the blood, 
but occasional escape from its natural effects has been recorded as above. 
It is quite as capable of causing emaciation as anemia. Interference with 
ingestion of food leads to anemia and emaciation in cancer of esophagus and 
stomach. The absorption of hemolytic agents from ulcerating and infected 
surfaces and from necrosing areas of closed tumors is a chief factor in pro- 
ducing anemia. Maragliano, Kullman, Bard, Polk, and many others have 
demonstrated the presence of hemolytic properties in the blood of cancer. 
Elsberg found that normal red cells injected beneath the skin of cancerous 
subjects were soon hemolyzed yielding a characteristic discoloration of the 
skin and he proposed this method as a diagnostic test for carcinoma. In 
extracts of degenerating and necrosing tumors Weil demonstrated thermola- 
bile and thermostabile hemolysins not differing from those obtained from 
necrosing normal organs. Chnical effects of the action of such hemolysins 
were observed by Bard who pointed out that in cancerous pleuritic exudates 
the red cells are often hemolyzed while in other bloody pleural fluids the 
red cells are intact. The presence of these circulating hemolysins may render 
transfusion of cancer patients with normal blood a dangerous procedure. 
Against their own serum the red cells are resistant and since the specific 
gravity of the blood is usually low the red cells are resistant to h}^otonic 
solutions. Since hemolytic agents and variations in the resistance of red 
cells are commonly present in other diseases these properties have not been 
found reliable in the diagnosis of cancer. 

The specific gravity of the blood in most cases does not differ from that of 
other forms of secondary anemia, but in well-established cachexia the gravity 
has often been found remarkably low. Dieballa, Peiper, and Moraczewski 
found lower specific gravity, loi 2-1032, in cancer of stomach, than in 
pernicious anemia of equal grade. The albumins of both plasma and serum 
are distinctly low in such cases (Grawitz). In a case of gastric cancer 
Wendelstadt and Bleibtreu found 0.79 gm. instead of the normal 2 to 2.25 
gms. in 100 gm. blood-serum. 

Rieder believed that the lymphocytosis of cancer resulted from accelerated 
flow of lymph and Grawitz attributed the dilution of the blood to the same 
factor, finding that watery extracts of cancer tissue when injected into 
rabbits produced a marked increase of lymph flow. 

Leukocytosis was early recognized as a very frequent condition in cancer, 
having been described post-mortem by Andral in 1823, and in the circulating 
blood by Lucke and Virchow about 1867. The leukocytes were long regarded 
as derivatives of the tumor. Virchow referred their presence to increased 
flow of lymph. 

Although some disturbance of the leukocytes is very frequently observed, 
the uncomplicated tumor process seems to have no capacity to attract 
leukocytes to the circulation. This conclusion accords with the fact that 
tumors excite no uniform local reaction. The cause of leukocytosis must 
therefore be sought in some other factor than the presence of growing tumor 

Ulceration or other inflammatory complications are the most definite 
causes of leukocytosis in cancer, and the resulting hemorrhage adds to the 
increase of white cells seen in bleeding and necrosing tumors. Bacterial in- 
fection, local and general, may intensify the effects of absorption of tissue 
toxins and loss of blood. 

Degeneration and necrosis of tissue and foci of cells must account for 


the leukocytosis of many large cellular and rapidly growing tumors. The 
inflained lymph nodes in the vicinity of malignant tumors indicate absorption 
of irritants but these nodes seem to furnish few new cells to the blood since 
the leukocytosis is usually polynuclear. 

As a reaction of immunity may be considered the lymphocytosis asso- 
ciated with certain tumors about whose edges round-cells gather in masses 
but there seems to be no ground in the human subject for assuming the same 
interpretation of polynuclear leukocytosis. With the onset of cachexia 
the way is open for the action of many causes of leukocytosis which are as 
numerous as the elements in the cachexia itself. 

With these theoretical principles the results of leukocyte counting in 
tumor patients offer only a partial parallel. The temptation is often strong 
to assume some specific chemotactic influence residing in the tumor process 
to account for pronounced leukocytosis observed in certain cases. Havem 
observed marked leukocytosis with scirrhus of the breast which subsided 
after operation only to return with recurrence of the growth. In many 
rapidly growing but apparently uncomphcated tumors especially in children, 
pronounced leukocytosis has been observed without satisfactory explanation. 
In the extensive series of blood reports of Alexander, Price-Jones, Cabot, 
and others, it is not always possible to refer the leukocytosis to definite factors 
apart from the tumor process. In fulminant carcinosis of breast and other 
organs the disease may be actively febrile and associated with marked poly- 
nuclear or mixed leukocytosis. Renal carcinoma is very frequently marked 
by intermittent fever and leukocytosis. Satisfactory explanation of these 
peculiarities is wanting. 

Certain specific events in the course of malignant tumors may be marked 
by definite changes in the leukocytes. Severe cachexia with anemia is usually 
attended with the presence of myelocytes, and myelocytes with nucleated 
red cells are often seen with marrow metastases. Kurpjuweit collected 
13 cases in which considerable numbers of myelocytes were associated 
with bone metastases. Eosinophilia is regarded by Neusser as diagnostic 
of bone-sarcoma. In general, sarcoma affects the blood and leukocytes earlier 
and more than carcinoma, and in many cases of sarcoma the increase has 
reached the higher limits of inflammatory leukocytosis (Martin, Matthew- 
son, Alexander). In certain cases of systemic lymphosarcoma, myeloma, 
and in leukemic chloroma, it is probable that numbers of tumor-cells reach 
the blood and may be found in stained specimens. In two cases of localized 
sarcoma Loeper and Loeste claim to have found specific tumor-cells in the 
centrifuged sediment of the blood. The many details of differential diagnosis 
by the blood may be left to the special treatises upon this subject. 

The total quantity of blood in cases of malignant tumor varies greatly but, 
as Louis showed in 1846, it is usually much diminished. The distinction 
between the anemic and the marantic cases is here quite pronounced, the 
vessels in the former type showing at autopsy a considerable volume of 
watery blood, while the others may give a much reduced quantity of con- 
centrated blood of fair quahty. The cases with arterial hypotension observed 
by Jane way must usually fall in the former class. 

The regeneration of the blood in cancer has been found by Bierfreund to 
require much longer than in other surgical conditions, and in progressive 
cases the Hb seldom reaches the ratio existing before operation. Yet it is 
impossible to establish any general rule in this matter. In a woman who in 
four years underwent six operations for recurrent adamantinoma of the 
superior maxilla, I found the highest Hb (85 per cent.) and good general 
health, although the tumor had again recurred. 


The resistance of the red cells to various destructive agents has been found 
about normal in early stages of cancer but in advanced cases it is much in- 
creased, probably as a result of reaction to hemolytic factors (Schmidlech- 
ner). So marked may the increased resistance become that Conte regarded 
this property as of diagnostic significance. 

Alkalescence of Blood. — Determinations of the alkalescence of the blood 
have yielded results which difier with the method employed and the particular 
factors brought into consideration. Klemperer found a reduced amount of 
CO2 in the blood of advanced cancer. Peiper titrating the whole blood 
found very low grades of alkalescence in advanced cachexia, and Rumpf, 
Limbeck, and others, dealing with the whole blood, obtained uniform dimi- 
nution of alkah in advanced cases. It is clear that the chief factor in these 
results was anemia, although as claimed by v. Noorden, liberation of acids 
from destruction of proteins and acidosis may occasionally be present 
(Herter). Yet H. Strauss titrating laked blood found marked variations, 
some cases showing normal and some increased alkalescence. 

Methods dealing with the serum yield directly opposite results. In 1906 
Moore and Wilson found that the serum alkalescence to dime thy lamido- 
azobenzol showed a striking increase in cancer as compared with healthy 
subjects and other hospital patients, while the basic capacity of the inorganic 
salts after removal of proteins by incineration showed a small but distinct 
increase. Gamble, Royle and Watson secured similar results by the same 
methods in both carcinoma and sarcoma, but in very advanced cachexia 
the alkalinity was barely above normal. Neither extirpation nor a-ray 
treatment produced any change. Sturroch improved the di-methyl tech- 
nic and found an average alkalinity of the serum of 0.190 N and often more 
than 0.200 N in cancer, while non-malignant cases gave an average of 0.173 N. 

Hyperglycemia with a percentage of glucose reaching 0.33 was regularly 
found in the blood of cancer by Freund and again in many cases by Trinkler. 
Yet Matrai could not find that the anemia of cancer varied in this respect 
from other forms. Lewis and Benedict found quite variable relations 
between sugar content of the blood and the progress of malignant tumors, 
but usually an increase in the later stages. 

A retention of chlorine and phosphorus, especially the former, was ob- 
served by Moraczewski to be earlier and more pronounced in the anemia of 
cancer than in other forms. 

Influence of Cancer on Digestion. — Gastric Digestion. — The absence of 
free HCl in the gastric contents of cancer of the stomach, discovered by Van 
der Velden in 1879, and thoroughly investigated by many later writers, 
stands as a prime factor in the anemia and malnutrition of this disease. In 
all but, 10-13 per cent, of such cases free hydrochloric acid is missing (Rich- 
ter). When cancer is grafted on simple ulcer the early stages of the process 
may be marked by the excess of free HCl belonging to the former condition, 
but as a rule this excess declines and eventually disappears, sometimes 
suddenly as the disease progresses (Rosenheim, Schneider). In most 
cases the loss occurs early in the disease (Riegel). 

Several factors appear to be combined in determining the loss of free HCl. 
In most cases of gastric cancer one finds a general chronic catarrhal inflamma- 
tion, in some cases there is atrophy of the mucosa, and in one group a diffuse 
infiltration of mucosa and muscularis is observed. There is every reason 
to attribute to these anatomical lesions a loss of functional capacity partly 
accounting for the loss of HCl. Yet the failure of free acid occurs in many 
early cases, and is relieved after excision of the cancer (Rosenheim). In 
other cancers not involving the stomach a similar deficiency of HCl occurs. 


SO that it seems necessary to assume that mahgnant disease exerts some gen- 
eral influence which affects gastric secretion, even in the absence of severe 
anemia and cachexia. 

Reissner showed that in gastric cancer the total chlorides secreted are 
not below normal, that decrease of free acid is nearly always associated with 
increase of fixed chlorides, and that the acid is promptly combined in the 
stomach with alkalis of the blood and serum exuded from the tumor. Since 
an increased supply of NaCl from the tumor juices can probably be elimi- 
nated, it would appear that the loss of free acid in early cases results from 
its neutralization by alkalis and not from any functional deficiency of an 
apparently normal mucosa. This explanation can hardly apply to late and 
cachectic cases and to other forms of cancer in which the total chlorides are 
diminished. That the tumor yields such free alkahs is further indicated by 
Stahelin's observation that free HCl introduced into the cancerous stomach is 
actively neutralized; by the neutralization of HCl by cancer tissue in artificial 
digestive mixtures (Rosenberger) ; and by the increased quantity of protein 
found in the stomach juice after previous washing (Salomon). 

Moore and Palmer from the study of a series of cases of carcinoma, 
sarcoma, and non-malignant tumors, found in nearly all a diminution of HCl, 
which they attribute to reduction of hydrogen ions of the blood. In many 
of these cases it is evident that they have to deal with the effects of cancerous 
cachexia, but in others which were not cachectic Parker suggests that the 
general unhygienic hospital surroundings were at fault. Copeman and Hake 
with Cramer, secured the interesting result from observations on 500 mice 
that the implantation of cancer in these animals is followed by decided 
increase in the physiologically active HCl. 

There are therefore a specific influence of tumor juices neutralizing 
secreted HCl of the gastric juice, a diminished functional capacity of an organ 
damaged by an extensive tumor, and an effect of cachexia in reducing the 
hydrogen ions of the blood and impairing the secretion of acid, all to be 
considered in the relation of cancer to the acid of the gastric juice. 

The gastric ferments persist long after the disappearance of free HCl 
(Oppler) . According to Glassner both rennet and pepsin are diminished with 
tumors of the fundus in which there is considerable destruction of the mucosa, 
whereas pepsin alone is diminished with cancers of the pylorus in which 
region it is principally secreted. Emerson found that one hour after a test 
breakfast 16.9 per cent, of the proteins in the normal stomach has been hydro- 
lyzed to the stage in which they are not precipitable by phosphotungstic acid, 
while in cancer of the stomach 27.6 per cent, reached this stage. In a -diges- 
tion mixture containing a piece of fresh cancer tissue he found proteolytic 
digestion to proceed further than in mixtures of pepsin — HCl. These 
experiments indicate that cancer tissue contains a ferment similar to that 
found in other autolyzing tissues. Whether this ferment is specific of 
cancer tissue; to what extent the advanced proteolysis observed in cancerous 
stomachs is the effect of this ferment; and the part played by leukocytes and 
bacteria, are matters yet to be elucidated. Rosenberg has verified these 
results and the method has been found of value in the diagnosis of gastric 
■ cancer. 

Neubauer and H. Fischer have found in the stomach washings of gastric 
cancer and in the expressed juice of carcinoma and sarcoma a ferment which 
spUts glycyl- tryptophan and this test has been extensively employed in diag- 
nosis with satisfactory results. Lately it appears to have been shown that this 
ferment is ptyalin from saliva. Fischer points out that amino-acids appearing 
in the abnormal digestion of cancer of stomach unite with free HCl through 


the amino-group, while the total acidity may remain normal on account of 
the acid reaction of the free carboxyl group. 

Disturbance of motihty of the stomach dominates the clinical signs in 
cancer of the pylorus, leading in many cases to dilatation of the stomach and 
stagnation with abnormal fermentations of the contents. Of the influence 
of cancer on absorption by the stomach little is definitely known. With 
diffuse infiltration of the wall and contraction of the cavity, evacuation of the 
organ may be accelerated and absorption must be greatly reduced. 

As a result of diminished digestive capacity, disturbance of motility, 
and retarded absorption, abnormal fermentative processes become estabhshed 
and various bacterial species find a favorable soil. The products of fer- 
mentation include lactic, butyric, and acetic acids, alcohol, and protein 
decomposition products. Boas at one time claimed that the presence of 
lactic acid in the gastric contents was specific of carcinoma and appeared at 
an early stage when stagnation and loss of free HCl did not exist, but it has 
since been shown that the presence of lactic acid is dependent upon impaired 
motility and deficiency of HCl (Strauss, Wagner, Seehg). Since these 
conditions are present very early in cancer and since, as Sick has shown, 
formation of lactic acid is favored by decomposition products of cancer 
tissue, the test for lactic acid is of considerable diagnostic value. 

Intestinal digestion suffers in many ways from malignant tumors of the 
gastro-intestinal tract and its accessory glands. With gastric cancer the 
discharge of abnormal digestive putrefactive and bacterial products into the 
intestine tends to incite disorder throughout the entire course of digestion, 
and Wasbutski, Kast and others have shown that intestinal decomposition 
occurs especially in cases with loss of free HCl and active fermentation in 
the stomach. Rarely there is severe diarrhea in the terminal stages of 
such cases. The main effects of tumors of the intestinal tract are the result 
of stenosis and ulceration and either of these conditions in pronounced form 
has a prompt influence on general nutrition. Detailed studies in this field 
are meager. Excessive indicanuria tells of absorption of putrefactive prod- 
ucts in stenosis, and formation of fistulous tracts may comphcate and 
terminate the course without much influence on the extent of this absorption. 

Blood in the stools offers a means of diagnosis in ulcerating tumors, and 
constitutes a serious source of anemia. Acholic stools result from can- 
cerous stenosis of the coinmon duct, and impaired digestion of fats follows 
destruction of the pancreas and occlusion of its duct. 

Tumors of the liver, bile passages, pancreas, cecum, and rectum, produce 
various important clinical types of disease the full discussion of which may 
be left to works on special pathology. Here it may only be said that all 
of these conditions tend to produce cachexia through the combination of 
very numerous factors which must be analyzed for each case and the variety 
of which reveals the very complex nature of tumor cachexia. 

Changes in the Urine. — Specific alterations in the urine of cancer have not 
been demonstrated, but its composition varies greatly according to general 
rules. Specific substances appear in the urine with melanoma and myeloma. 
The entire subject invites further study by improved methods. 

The total nitrogen varies with the diet throughout the main course of the 
disease, while at the termination may be observed the increased nitrogen 
output which Miiller interpreted as a sign of toxic destruction of proteins, 
or the diminished excretion of starvation. The distribution of the urinary 
nitrogen fails to show any specific characters. The percentage of urea is 
influenced mainly by starvation, falling to a very low figure in extreme cases. 

Uric acid may be increased without relation to leukocytosis (Carlo). 


The course of this substance in cases of various cellular tumors has not been 
determined. Brandenburg and Blumenthal found the nitrogen of alloxur 
bodies relatively high in most of their cases, while Setti failed to find any 
uniform increase. 

Ammonia nitrogen has not been found to show any variations other than 
those dependent on digestion and nutrition. Herter found high ammonia 
in the coma of gastric cancer. 

Creatinin excretion has been found low in cachectic cases by Benedict 
and Gorslin, an observation which accords with Shaffer's view that the 
creatinin coefficient is a measure of muscular efficiency. Creatin appears 
in moderate amounts in cachexia. 

Rest nitrogen, chiefly amino-acids, is higher than in some other abnormal 
states and increases with the progress of the disease (,Setti). Salkowski has 
drawn attention to an increase in the colloidal nitrogen precipitable by alco- 
hol, in cancer. This fraction corresponds partly to the amino-acids (Wolf). 
Later Salkowski estimated the nitrogen precipitable by lead acetate after 
removal of phosphates by alkaline barium chloride, finding that this nitrogen 
fraction is greatly increased in cancer. The nitrogen estimated by Salkowski 
probably includes the oxjrproteic acids and polypeptids. 

An increase in unoxidized sulphur has also been observed by M. Weiss 
and others, and Salomon and Saxl have demonstrated such a notable and 
uniform increase in oxidizable neutral sulphur as to indicate its value in 
diagnosis. Lehman points out that the results of this test depend largely 
on the diet without the control of which the examination is useless. From 
observations in this field Saxl concludes that various analytic methods 
demonstrate the occurrence in notably increased amounts of protein deriva- 
tives, chiefly oxyproteic acids, which escape the normal transformation 
into urea. In general he finds a somewhat specific disturbance of metab- 
olism in cancer marked by a nearly constant excess in total metabolism, 
relative reduction in urea formation, increased ammonia excretion and 
increase of protein derivatives which fail of oxidation into urea. In no 
other disease except certain special intoxications is the internal oxidation of 
protein so much disturbed as in cancer. Saxl goes on to show that the 
disturbance of oxidation is probably due to accumulation in the system of 
rhodan an oxidation product of hydrocyanic acid, since this substance is 
increased in the urine in cancer and when administered to man reproduces 
the typical metabolic disturbance of cancer. The basis of this theory remains 
to be verified. 

The urinary chlorides show no specific relation to cancer. Although 
Beneke supposed that the cancerous constitution was marked by deficiency 
of alkaline chlorides, observation shows that these urinarj^ principles depend, 
as in other diseases, chiefly on the diet, the presence of disintegrating protein 
tissues, and the activity of the kidneys. In severe inanition the chlorides 
fall very low (Miiller, Braunstein), but an increased excretion has often been 
observed (Schopp). Chloride retention may occur with cardiac or renal 
insufficiency and may be associated with anemia and edema (Laudenheimer). 
Chlorine loss may result from active destruction of tumor or other tissues. 
Marked disproportion between urinary nitrogen and chlorine in favor of the 
latter occurs in states of inanition with Httle protein intake and much destruc- 
tion of tissue, as in cancerous stenosis of esophagus. Schopp concluded that 
NaCl might be retained in the building up of tumor-tissue in cancer of the 
liver, but Laudenheimer found the same proportion of NaCl in liver tumor 
and normal hver tissue. Royle finds retention of both Na and CI in cancer, 
that of Na being greater than CI. He concludes that extra Na combines with 


phosphoric acid and that the retention of chlorine seems to counterbalance 
the alkahne phosphates of the blood and maintain the proper relation of 
salts in the plasma. In lo cases of carcinoma Braunstein found no changes 
in the urine which could be regarded as specific. In three cases the excretion 
of Na and CI was practically norrrial and the loss of P ran parallel with that of 
N. Phosphoric acid excretion is increased in most cases both absolutely 
and in proportion to the nitrogen with which in general it runs parallel. The 
P : N ratio runs i : 4 or i : 5 : 6 instead of the normal i : 7 (Braunstein) . Its ex- 
cess suggests a source in tissues rich in P, as the bones, glandular organs, or the 
tumor tissue itself. Sulphuric acid excretion is parallel with that of nitrogen 
but Carlo observed somewhat increased proportions in two cases of cancer of 
esophagus. The partition of the sulphur has not been determined, but high 
percentages of ethereal sulphates may be expected with ulcerating tumors, and 
intestinal disturbance. 

Urobilin. — Considerable urobilin was frequently observed and regarded 
as somewhat characteristic of cancer especially of cachectic and hepatic cases, 
by F. Miiller and Gerhardt. Yet Hoppe-Seyler had previously found little 
urobilin in several cases, and Blumenthal who observed an excess in gastro- 
intestinal and ulcerating cases, often found it absent in others. Braunstein 
from the study of 22 cases concludes that urobihn is missing in cancer so 
long as fever and complications are absent, but appears with metastases 
especially in the liver; with necrosis and decomposition of the tumor; and 
with complicating pneumonia. In cancer of the liver when the bile ducts 
are closed he found that bilirubin replaced the urobilin. 

Products of putrefaction of proteins are found in increased quantities 
in the urine of advanced cancer but the exact source of these substances has 
not been fully determined. Indican, phenol, aromatic oxyacids, and ethereal 
sulphates, have been estimated by several observers who attribute their 
presence to intestinal putrefaction, to decomposition of secretions and to 
destruction of tumor-tissue. That intestinal putrefaction is not the sole 
source is indicated by the considerable quantities of such substances observed 
with decomposing tumors of breast and uterus (Brieger, HaberUn, Hennige). 
Lewin finds excess of aromatic acids in patients with N loss, but not with 
those showing N retention and concludes that these substances are derived 
from the toxic destruction of tumor and tissue proteins and not wholly from 
intestinal putrefaction. Yet it still remains uncertain whether aromatic 
oxyacids, indol and phenol, may arise in the course of nitrogenous metabolism, 
and it is clear that patients laying on nitrogen are not likely to be suffering 
from excessive intestinal disturbance. 

A cidosis of Cancer. — The acetone bodies in cancer follow the same rules 
as in other conditions. Their presence in the urine bears no relation to the 
growth of cancer as such, and these substances are absent in the early stages 
of the disease and when the patient is well fed (Waldvogel) . With the 
advent of protein loss and cachexia acetone and diacetic acid commonly 
appear in the urine and beta-oxybutyric acid is added in severe inanition. 
Hirschfeld relieved the acidosis by giving sugar. 

In cases chiefly of esophageal, gastric and hepatic cancer v. Jaksch, 
Riess and Senator described a form of cancer coma in which the dyspnea, 
acidosis, and diminished alkalescence of the blood were present as in diabetes. 
In two such cases Klemperer observed a daily N loss of 5 to 9 gms. which he 
interpreted as signifying toxic destruction of tissues. The coma he referred 
not to the acidosis but to the general toxic process, v. Jaksch also stated that 
these patients were not always suffering from starvation but that the appear- 
ance of acetone bodies ushered in a severe type of general cachexia. It is not 


entirely clear that cancer coma is purely an acid intoxication, and it seems 
probable that the rapid bu'rning of tissue fats and proteins of which acidosis 
is only one result may be responsible for some of the sudden terminations of 
malignant tumors. 

Von Noorden's observation of considerable quantities of lactic acid in the 
urine of two cases suggests that the acidosis of cancer may involve hepatic 
disturbance as well as pure acidemia. 

Demineralizalion occurs in cancer cachexia as in other conditions and is 
associated with nitrogen loss with which it maintains a parallel. Of ii 
cases Lewin observed loss of mineral salts, 1.36 to 8.9 gms., and of nitrogen, in 
seven, and slight retention of both in four. One important source of these 
mineral salts is probably the bones which become increasingly fragile in 
many wasting diseases. The shafts become thinner, the cancellous spaces 
and Haversian canals larger, and Campbell has shown that the breaking strain 
is greatly reduced. In some cases of cancer the calcium depots become 
greatly disturbed and extensive deposits of calcium salts are found frequently 
in the tumor, occasionally in the lung. 

Albuminuria occurs in most cases of cancerous cachexia but is absent in the 
early stages of the disease. According to Miiller it occurs in 35-72 per cent, 
of all cases of carcinoma. In cancer of the intestinal tract albumoses are 
very often present (Ury, Lilienthal), either by direct absorption of digestive 
products through the ulcerated surface (Maixner) or from the disintegration 
of tumor tissues (Pacanowski). In many cases of myeloma a peculiar pro- 
tein, Bence- Jones body, appears in the urine. 


The formation of secondary tumors is a cardinal property of maKgnant 
tumor growth. The appearance of a secondary tumor at some distance 
from the primary growth may represent one of several events. It may mean 
a multiple origin, and a new tumor; or a continuous extension from the 
primary growth; or a true metastatic growth arising from tumor-cell emboli. 
Only a critical estimate of the known tendencies of particular tumors can 
determine which of these events has occurred. Melanoma and multiple 
fibroma both produce very numerous tumors of the skin, the one from cell 
emboli, the other by multiple growth of new tumors. 

Multiple origin of new tumors is observed with a considerable list of 
neoplasms, already considered, including fibroma moUuscum, myeloma, 
epithehoma of skin, sarcoid growths of skin, adenoma and adenocarcinoma 
of mucous membranes, chondroma of spine, myoma, hpoma, etc. 

Lymphatic Permeation. — By continuous extension of primary infiltrating 
tumors through the lymphatics, apparently- discrete tumors often appear 
at points widely distant from the original growth. In multiple carcinoma 
of the skin continuous chains of cells may extend through convenient lym- 
phatics until the cells reach a locality favorable for growth and there mul- 
tiply, producing discrete tumors connected with the primary growth only 
by a thin strand of cells. It is conceivable that accidental breaks may occur 
in this chain of cells completely separating the secondary from the primary 
growth. There are many tumors in which this process leads to formation 
of secondary growths. At what remarkable distances secondary tumors may 
arise by this method has only recently been demonstrated. 

In cancer of the breast Handley has shown that there may be continuous 
extension of tumor-cells through the lymphatics of the deep pectoral fascia 
to the axillary and supraclavicular nodes: thence to overlying skin and to 
humerus: through the deep lymphatics to the ribs, pleura, lung and spine: 
across the chest wall to the opposite breast: downward through the abdom- 
inal wall to the epigastric region and thence by the falciform ligament to the 
liver; and further to the inguinal region with involvement of lymph-nodes, 
skin and femur. Extension through superficial lymphatics is much less 
wide. Invasion of the humerus occurs at the deltoid insertion and of femur 
at the trochanter, where these bones are closest to the skin, through the 
deep lymphatics of which the bone invasion follows. The leg and arm bones 
escape infection, which is inconsistent with a free embolic origin through the 
blood-vessels. Passing through the deep lymphatics of chest or abdominal 
wall the cells enter the pleura and peritoneum and become implanted on 
the serous surfaces and produce superficial infiltrations of the lung, liver, 
intestines, and ovary. Or the viscera may be invaded through their main 
lymphatic vessels giving central tumors. The Hver is involved through the 
lymphatics of the falciform and round Ugaments, or by transperitoneal im- 
plantation, or by way of portal nodes; the lung by transpleural implanta- 
tion or through bronchial nodes and hilus. Abdominal invasion by the epi- 
gastric route is earlier and more frequent than the thoracic, occurring without 



thoracic lesions in 12 per cent, of all cases. The diaphragm is invaded by 
the epigastric peritoneal route, and through the descending lymphatics of 
its crura the retroperitoneal nodes and kidneys are attacked. This process 
of lymphatic permeation Handley believes is the master process of general 
dissemination in cancer. Lymphatic embolism he would reduce to a very 
secondary factor. Handley bases his conclusions on the study of long 
sHces of thoracic and abdominal parietes in advanced cases of mammary 
cancer. In such material he finds the lymphatics continuously filled by 
tumor-cells. The only breaks occur as the result of perilymphatic fibrosis, 
an inflammatory reaction excited by the tumor process and resulting in 
complete atrophy of tumor-cell cords at many points. 

While recognizing the value of Handley's studies it must be considered 
that the condition at autospy may not clearly indicate the manner in which 
the lymphatics have become filled. It is conceivable that cell emboli lodging 
at different points in a lymphatic chain may grow in both directions and 
after a time fiU the lymphatic completely. Many observers believe that 
lymphatic embolism is commonly associated with continuous permeation 
in most mammary and in many other cancers. However, the result of these 
processes is practically the same, ending, in prolonged cases, in very wide- 
spread and nearly continuous invasion of lymphatics. 

The theory of continuous permeation is of fundamental importance in 
surgical procedure. To what extent this process prevails in different forms 
of cancer can be determined only by careful study in each case. Such 
:3tudies are not yet available and are essential before this theory can safely 
be adopted. In epithelioma of esophagus Borrmann has found that the 
metastatic tumors in the stomach wall, which have long been regarded 
as implantations through the esophagus, are connected with the primary 
tumor by an unbroken chain of cancerous lymphatics. Bloodgood believes 
that successful extirpation of cancer of the tongue and infected lymph-nodes 
may be accomplished only by removal of all lymphatics joining ulcer and 

Yet the majority of observers fail to find in early cases such uniform and 
continuous connection between primary cancers and infected regional lymph- 
nodes and are agreed that such infection is usually by means of cell emboli. 
I have repeatedly failed to find any trace of invaded lymphatics connecting 
carcinomas in the breast with infected axillary nodes, and have been equally 
unsuccessful with cancer of tongue and penis. Metastases from Ungual 
cancer often appear first in rather distant nodes, sometimes on the opposite 
side of the neck. The metastases from melanoma of foot and other regions 
are usually incompatible with the theory of continuous permeation. In 
general it appears probable that the rapidly growing epidermoid and glandu- 
lar carcinomas disseminate chiefly by lymphatic embolism, while slowly 
growing and recurrent tumors, especially in the skin often extend by continu- 
ous permeation. This belief is strengthened by such observations as those 
of Gussenbauer who found cancer-cells in the cervical lymph-nodes in 29 
of 32 cases, and of Kuster who found the axillary nodes free from cancer-cells 
in only two of 117 operative cases of breast cancer. Since definite growing 
tumors of the lymph-nodes are much less common it is clear that embolic 
cancer-cells are frequently present in lymph-nodes and that they are prob- 
ably destroyed there under some circumstances. Yet it is extremely rare 
to find in lymph-nodes any appearance suggesting the destruction of sniall 
groups of embolic tumor-cells. On the other hand evidences of retardation 
of growth by fibrosis and encapsulation are relatively common. 

The frequency of lymph-node metastases varies with different tumors. 


with different varieties of the same tumor, and with different positions of the 
tumor in the same organ, and must be separately considered for each case. 

Cancer of breast probably stands high on the hst, very few fully mahg- 
nant cases of this disease faiHng to involve the nodes if allowed to remain 
more than a few months. Williams found the nodes involved in 86 (73 per 
cent.) of 118 cases at first examination, and in 40 (90 per cent.) of 44 autop- 
sies. The variety of the cancer is here of paramount importance, and is 
a factor often underestimated. 

Epithehoma of the tongue or tonsil very early and nearly constantly 
invades the lymph-nodes but here again the type of the tumor is the deter- 
mining factor. WilUams recorded 86 (83 per cent.) of 104 lingual cases 
infected at time of first examination, and 56 of 57 cases coming to autopsy. 
Epithehoma of the lower lip invades the lymph-nodes more slowly but 
eventually in almost all cases. In 186 cases o^ cancer of stomach exa,mined 
at autopsy by Moore, Cuneo, and Colwell, lymph-nodes were involved in 149. 

The lymph-nodes escape in many cases of uterine cancer, especially with 
tumors of the fundus. Epithelioma of the skin is slow in reaching the lymph- 
nodes, and rodent ulcer and adenoid cystic epithehoma are remarkable 
for the long or permanent immunity of adjacent lymph-nodes. Of 34 malig- 
nant tumors of the testis Butlin found the lymph-nodes free in only three, 
but in a series of 19 cases I found several which had not attacked the inguinal 
nodes. The tumor usually first affects the epigastric nodes. 

The value of available statistics, however, is very limited since most 
authors fail to distinguish between the different types of cancer, although 
it is well known that some are very prone to invade the lymphatics while 
others long remain circumscribed. It is httle help to know that 75 per cent, 
of the cancers of the breast involve the lymph-nodes within a year of their on- 
set when the surgeon might learn that his case is an adenocarcinoma which 
long spares the nodes. 

For the same reasons available statistics regarding the period at which 
lymph-node invasion occurs appear to be of uncertain value. Thus Fink 
observed it in breast cancer as early as the sixth and always after the thir- 
teenth months, yet I have examined a mucoid breast cancer of 10 years' 
duration, service of Dr. Hartley at Roosevelt Hospital, without invasion of 
the nodes. Similar variations are observed with many other tumors, while 
it should be noted that some cancers show general dissemination without 
palpable enlargement of regional nodes. 

Changes in lymph-nodes draining mahgnant tumors show that the im- 
plantation of metastases is preceded by a period of preparation of the soil. 
For weeks or months before actual tumor invasion the regional lymph-nodes 
may be moderately swollen. During this period many new lymph-nodes may 
develop in the course of the vessels. On section the swollen nodes show 
.diffuse hyperplasia with catarrhal exfoliation of sinus endothelium, or multi- 
plication of follicles. It is not uncommon also to find the nodes atrophic 
and- fibrous, or extensively invaded by fat tissue, conditions which reduce 
their effectiveness as filters. Owing to a variety of causes old changes of 
this latter class may permit the passage of cancer-cells through or around 
a given group only to lodge in a more central area. The former class of recent 
alterations must be referred to the absorption of toxic products from the 
tumor, autolytic and bacterial. There are no specific histological features 
of the precancerous condition of the lymph-nodes, but there are reasons to 
believe that regressing tumor-cells may be found in the sinuses before definite 
implantation has occurred. As previously noted, Kuster found that in 
only two of 117 cases were the axillary nodes removed with breast cancer 


free from invasion. Yet the proportion of lympli-node recurrences when the 
nodes are not removed is very much less. Nevertheless it is extremely rare 
to find any metastatic focus in a lymph node which does not appear to be 
safely implanted and capable of growth. 

The earliest nodules appear in the sinuses from which they invade and 
compress the pulp with complete atrophy of lymphoid tissue. From this 
point, the cells may invade the capsule and the peripheral lymphatics, 
blood-vessels, fat, or fibrous tissue. 

When the metastatic period is established lymph-node invasion may 
follow rapidly. In a case of cancer of breast a swollen axillary node examined 
in March showed only inflammatory hyperplasia with no demonstrable 
tumor in the breast, but in June the entire breast was infiltrated by a flat 
diffuse growth and some axillary nodes were completely replaced by cancer. 
Since many tumor metastases long remain confined to the regional lymph- 
nodes and fail to make any headway through the blood-stream it may be 
assumed that there is a regional and a general immunity against implantation. 
Yet in not a few cases cancer skips the regional nodes and yields distant 
metastases. According to Gross this occurred in one out of 7 cases of 
breast cancer. With some rapidly growing tumors both local and general 
susceptibility to metastasis exists at a very early period, and in most cases 
one must conclude that the growth capacities of the cells are of chief impor- 
tance in determining the fate of cell emboli. 

Invasion of the Thoracic Duct. — In not a few cases of advanced cancer 
the thoracic duct is invaded, an event which commonly leads to widespread 
dissemination by lymph and blood-stream. 

One of the striking results of invasion of the thoracic duct is the appear- 
ance of enlarged nodes in the supraclavicular region. Troisier in 1886 
first emphasized the diagnostic importance of this symptom in cases of 
abdominal and especially of gastric and uterine cancer, and similar cases 
have since been reported by many observers (Lit. in Williams, p. 425). 

It has also been observed with cancer of pancreas (Raw), testis (Poncet), 
prostate (Hurlemont), adrenal • (Troisier) . Nevertheless it is a late and, 
with improving diagnosis, a rare symptom of diminishing importance, and 
must be distinguished from other affections of the nodes, and according to 
Wilhams, from Verneuil's "pseudolipome sus-claviculare." According to 
Rousseau, in 37 cases presenting Troisier's symptom, the left nodes were 
involved in 29, the right in four, both sides in four. 

Inguinal lymph-node invasion is also associated with and occurs without 
involvement of the thoracic duct, with abdominal cancers, especially of 
the uterine fundus, prostate (Viannay), bladder, rectum, chorion, and occa- 
sionally with many- other primary tumors. Among other symptoms resulting 
from occlusion of the thoracic duct are chylous ascites (Leydbecker), chylo- 
thorax, and lymph varices which may lead to cysts of considerable size. The 
results of invasion of the blood-stream from the duct with multiple tumors 
of the lung are frequently observed. 

Winkler has given a full description of 27 cases (12 original) of cancer and 
one of round-cell sarcoma of the thoracic duct. The invasion followed most 
frequently cancer of the stomach, then of uterus, while other cases arose 
from the gall-bladder, testis, colon, and kidney. In all these cases the abdom- 
inal nodes were enlarged and adherent to the duct, offering abundant oppor- 
tunity for the invasion by tliis means. The duct presented one of three 
conditions, (i) Free masses of tumor-cells lying in a coagulum, with throm- 
bosis or occlusion (2) Papillary masses of tumor growth adherent to the wall 
and organized from it, usually at the valves, with numerous varices often 


resembling a string of pearls. (3) Complete thrombosis by tumor masses 
invading the wall and mingled with fibrin. 

Invasion of Serous Cavities. — The large serous cavities are penetrated 
by many pathways and in the course of a considerable number of tumors 
chiefly those of the contained organs. The most frequent source of free 
peritoneal growths is the ovarian adenocarcinoma which at any period may 
rupture its covering and be disseminated throughout the cavity. Its cells 
readily become implanted on the peritoneum, producing many miliary or 
large sohd papillary or cystic tumors with a tendency to ascites. One of 
the most remarkable forms of peritoneal cancer arises from small adeno- 
carcinomas of colon or appendix, which early perforate the peritoneum, 
spread rapidly throughout the greater sac and, retaining the large alveolar 
structure, produce enormous quantities of mucus greatly distending the 
abdomen. In very large tumors of this sort, several liters of gelatinous 
material may be evacuated and it may be extremely difficult to discover any 
trace of living tumor-cells. I have seen gelatinous cancer of the rectum, 
following an extraperitoneal route, fill the pelvis inclosing rectum and bladder, 
pass up the abdominal parietes in front and retroperitoneal tissues behind, 
separating liver from diaphragm, inclosing the entire abdominal cavity in 
a rigid shell 3-5 cm. in thickness, and eventually invading the mesentery 
and subserous intestinal lymphatics, but without any trace of intraperitoneal 
growth. This condition may be designated as abdominal cancer en cuirasse. 

Peritoneal invasion is a very common late complication of breast cancer, 
and occurs as Handley has shown, often by way of the epigastric lymphatics. 
From the subperitoneal lymphatics and along the falciform and round liga- 
ments of the liver he has found cancer nodules opening into the peritoneum. 
Loose cells often become implanted first in Douglas sac or on the ovary. 
More frequently there is a widespread permeation of the subperitoneal 
lymphatics, which may become nearly universal. It is associated with active 
growth of hard fibrous tissue, producing nodules and adhesions. Eventually 
the entire peritoneum becomes thickened, leathery and opaque, and the 
omentum shrinks to a firm globular infiltrated mass. Such patients show 
marked resistance to cachexia and may suffer for months or years from increas- 
ing ascites which gradually drains their strength or suddenly terminates 
life by asphyxia. 

The peritoneum seems to have very little capacity to destroy cancer- 
cells, so that the invasion once accomphshed is usually progressive. Gravity 
and the muscular contraction of stomach, intestine, diaphragm and parietes 
assure a general dissemination of cells. These movements also may cause the 
separation of tumor masses which have been found free in the peritoneal and 
pleural cavities, while the aspirated fiuid nearly always • contains exfoliated 

Intrapleural and intrapericardial cancerous invasion occurs chiefly by 
direct invasion from cancer of breast, bronchi, and lung, with results very 
similar to those produced by breast cancer in the peritoneum. 

Penetration of the cerebrospinal membranes occurs in rare cases of carci- 
noma chiefly <5f the breast, and may lead to a universal cerebrospinal carci- 
nomatosis resembhng meningitis. In a breast case observed at New York 
Hospital, service of Dr. Conner, isolated cells, small groups of cells and thin 
plates were found from the vertex to the cauda equina, infiltrating the pia, 
and penetrating many sulci, with little cellular or serous exudate. 

Lymphatic Invasion by Sarcoma.— It is a somewhat important dis- 
tinction that lymphatic invasion by carcinoma is frequent, by sarcoma rare. 
Not a few epithelial tumors violate this rule and travel through the blood- 


vessels, but sarcomas that tend to invade lymphatics are much less numerous. 
The lymphosarcomas are the only sarcomas that frequently travel by the 
lymphatics and the origin of these tumors within the lymph-nodes fully 
accounts for this tendency. The blood-vessels also serve as channels for the 
dissemination of lymphosarcoma, so that certain forms of this tumor yield 
the most abundant and widely disseminated metastases seen with any type of 

Most statistical reports of lymph-node invasion in sarcoma are of doubtful 
value owing to the uncertainty connected with the diagnosis of sarcoma. 
Most of the sarcomas reported of the testis and thyroid are probably epithe- 
lial tumors, and commonly invade the lymph-nodes. Melanoma declares 
its specific nature by frequently (41.7 per cent. Williams) affecting lymph- 
nodes. Of the bone-sarcomas the spindle-cell and giant-cell types rarely, 
the round-cell or lymphoid-cell types frequently involve the adjacent nodes 
(Gross). The frank spindle-cell sarcomas of the parietes and its contained 
structures, and those of the internal organs affect lymph-nodes in 5-15 per 
cent, of cases. As a rule the blood-vessels as well as the lymphatics may be 
the route of entry. 

The comparative immunity of lymphatics against invasion of sarcoma 
is probably to be explained chiefly by the greater local fixation of sarcoma- 
cells as compared with the mobile and sometimes amebiod cancer-cell. 
It is in accordance with this view that the lymphosarcomas behave much 
like carcinomas. At the same time sarcomas are usually more vascular 
than carcinomas so that cells rapidly break into the blood-vessels whose 
walls are thin or deficient. The growth of most sarcomas is more expansive 
than with carcinomas and more often encapsulated, so that lymphatic 
channels are closed and only the nutrient blood-vessels remain pervious. 
I do not think that the original relations between lymphatics and epithelium 
or connective tissue cells have any influence on the character of tumor metas- 
tases, since at the period when dissemination begins these relations are 
greatly altered. 

Dissemination by Blood-vessels. — The blood-vessels are the chief 
channel of the extension of sarcomas, they are frequently invaded by carci- 
nomas, and several epithelial tumors show a remarkable capacity to flourish 
in and travel through the veins. The abundance of thin-walled blood-vessels, 
the presence of naked sinuses, and the tendency toward necrosis and hemor- 
rhage rather than fibrosis, may largely explain the predominance of blood- 
vessel invasion by sarcoma as compared to carcinoma. In many cases, as 
in fascial myxosarcoma, the tumor-cells appear to be mechanically forced into 
the veins by the presence of rapidly growing and swollen tumor-tissue. 

The entry into the veins is accomplished through the defective vessels 
within the tumor, or by invasion and rupture of adjoining vessels, or by infil- 
tration along the vasa vasorum and intima, with gradual erosion of the wall. 
Carcinoma often penetrates the wall via the perivascular lymphatics. 

The result of such invasion is at first a thrombosis complete or partial, 
at the point of entry. The tumor-cells may then be destroyed by organi- 
zation and fibrosis of the thrombus with subsequent canalization (Schmidt), 
or the cells continue to grow within the vein, they become attached to the 
wall, produce flat or polypoid outgrowths, many cells or cell masses become 
detached from the wall or from the original tumor and are carried to the 
first zone of capillary vessels, usually the lungs, where they lodge as emboh. 
Here again they are exposed to destructive agencies and may perish, but 
more often develop secondary tumors. 

If the detached masses are large they regularly lodge in the lungs, or 


become arrested in the right heart. Direct transport of emboli or continuous 
growth in the veins is by far the most important mode of dissemination in 
blood-vessels, and several well known tumors offer striking illustrations. 
Chorioma originates in the uterine sinuses and maintains throughout its 
intravascular and hemorrhagic tendencies, producing retrograde vaginal and 
renal metastases, honeycombing the uterine wall, and yielding forwardmetas- 
tases in the lungs. Portions of the stroma of villi as well as tumor-cells have 
been found in the pulmonary emboli of chorioma. Hypernephroma regularly 
perforates the renal vein and may extend in a solid column of tumor-tissue 
into the heart. Adenoma of the Uver is commonly found distending the 
portal veins. The intravascular hydatid growths of teratoma testis are 
among the oldest observed forms of intravenous extension. When a sarcoma 
or carcinoma ruptures into a pulmonary vein widespread dissemination 
results and very numerous tumors appear rapidly in different tissues and 

The form assumed by tumors growing within vessels is sometimes influ- 
enced by the existing mechanical and nutritional conditions. Freedom from 
pressure allows the cells to reveal their natural tendencies, and the intra- 
vascular tumor may be more "typical" than the original. Or a more rapid 
and diffuse growth is assumed. In the axillary vein the cells of breast 
cancer may become larger and more anaplastic, and I have seen evidence of 
multiplication of loose cells in this situation and in the thoracic duct. The 
organization of a tumor thrombus with fibrosis and reopening of the lumen 
occurs rarely. The advancing mass of hypernephroma in the renal vein is 
usually very edematous and the cells hydropic. Fuhrmann and Risel beheve 
that some tumors, especially myxoma, tend to assume an hydatidiform growth 
when invading blood-vessels and they would attribute the chorion-like form 
of certain testicular teratomas to the influence of the intravascular position 
and not to any intrinsic chorionic character. 

Are vagrant tumor-cells destroyed in the blood or lymph stream or their 
associated organs, and if so by what mechanism? To this fundamental 
question only a very incomplete answer can be given. 

The scope of the conditions to which the question applies is very wide, 
including many different forms of tumors and cells, involving the natural 
hfe span of the cells and the facility with which they may reach the circu- 
lation. It must be assumed at once that in many systemic sarcomas and 
other rapidly growing neoplasms, cells are constantly reaching the circulation 
and perishing there at least by natural autolytic processes. It is a very 
prevalent belief however, that cells of most carcinomas from an early period 
of their growth frequently reach the blood and lymph and are actively 
dissolved by antagonistic substances in the blood. Many general considera- 
tions favor this view. The active infiltration of capillaries and lymphatics 
by rather loose tumor-cells suggests an easy and frequent passage into the 
moving current. The list of observed emboH even of normal cells is quite 
large. The experimental cultivation of tumor-cells by Burrows shows that 
they are regularly capable of growth in the isolated state, and in many human 
tumors, as in thyroid adenoma, it has been shown that blood-clot may be 
penetrated by loose tumor-cells much as in an artificial culture. In several 
well-known tumors cited elsewhere, as chorioma, hepatic carcinoma, and 
mammary carcinoma, tumor masses and loose tumor-cells are observed 
within the vessel lumen. 

The belief that such loose tumor-cells are destroyed in the blood- or 
lymph-stream is based on rather scanty evidence. Usually wide dissemina- 
tion is observed to follow the invasion of blood-vessels. M. Schmidt, 


however, has shown that tumor-cell emboli are mach more frequent in the 
lungs than are secondary tumors of these organs. In 28 cases he found 
emboli in the pulmonary arterioles in all, but definite metastases in only 
13. From the study of the lungs in these cases Schmidt concludes that 
in carcinoma of the abdominal organs there is frequently and repeatedly a 
discharge of cancer-cells which lodge in the small arteries of the lungs. 
Only a small portion of these cell emboli produce metastatic tumors or grow 
into the perivascular lymphatics. Most of them are inclosed in fibrin, 
organized, encapsulated, and in spite of retained vitality, rendered harmless. 
Some of them grow through their capsule into the capillaries and small 
veins whence they may form the source of late general metastases. All 
of these processes may occur without macroscopical changes in the lungs. 
The chief source of these emboli he believes to be through the pervious 
thoracic duct. The endophlebitis carcinomatosa which Goldmann finds to 
be nearly constant in generalized carcinoma, he regards as a sequel of pul- 
monary emboli. In chorio-adenoma it has been shown that pulmonary 
emboli occur in cases which recover after hysterectomy, but the mechanism 
of destruction of the embolic cells, if they exist, has not been followed. In 
many hepatomas the growth is from the first into the veins, but pulmonary 
or other metastases are almost unknown with this tumor. 

In the lymph-nodes while it has been shown that microscopic invasion 
is more frequent than growing metastases, actual observation of late regress- 
ing cell emboli in the nodes is not satisfactorily reported. The lymph-nodes 
seem to retain and retard the growth rather than destroy the cells. 

Experimental research has suggested the existence of a premetastatic 
period during which the organism is incapable of developing metastatic 
tumors. Gay found that he could not successfully reinoculate tumor-bearing 
animals until a certain period had elapsed after the first implantation. 
Yet these results are not uniform. Tyzzer by manipulating his grafted 
tumors was able to cause numerous metastases apparently without regard to 
the period of growth. 

Serological studies have in general failed to demonstrate in the blood 
serum of natural or immune animals any substances capable of destroying 
tumor-cells. The most direct evidence favoring the existence of cytolytic 
powers of the blood is contributed by Freund and Caminer. These observers 
claim to have shown that normal blood dissolves cancer-cells in vitro, while 
the blood of cancer patients has lost this property. With many others I have 
been unable to verify these observations, and believe that Freund was dealing 
with erratic autolytic and bacterial effects and not with true cytolytic agents. 
In any case Freund's studies would not encourage the view that tumor-cells 
are destroyed in the circulating blood of cancer patients. 

Retrograde transport was first emphasized by v. Reckhnghausen as an 
important process in the dissemination of tumors through both blood- and 
lymph-vessels. The basis of his conclusions had been laid in many previous 
observations on the fate of insoluble particles injected into the veins. Retro- 
grade transport occurs chiefly in those organs in which there is a normal 
venous pulse, lungs, liver, kidneys, heart, and brain, and in which violent 
expiration or increase of intrathoracic pressure from any cause may trans- 
form the weak forward into a backward current. When the main lymphatic 
or vein is occluded the flow is entirely disordered and backward transport 
is faciHtated. Occlusion of the trunk vessel is probably essential for back- 
ward metastasis in lymphatics but not in veins. In either case the backward 
progress is probably slow and intermittent although wide distances seem at 
times to be traversed (Arnold, Ernst). In the organs retrograde metas- 


tases arise from tumors originating in these organs and from distant tumors. 
Thus a large tumor thrombus lying in an arched vein of the kidney in primary 
sarcoma of tibia could originate only by backward passage through the renal 
vein of an embolus traveling from below. Emboli in a cerebral sinus with 
cancer of breast (Arnold), in an hepatic vein from intestinal or thyroid cancer 
(Heller, Bonome), and in cerebral, cardiac, pulmonary, and hepatic veins 
with hypernephroma ' (Ernst), are not infrequent examples. Backward 
passage through obstructed lymphatics may be assumed in cases of advanced 
cancer of bronchial nodes with subpleural lymphatic invasion but without 
involvement of pulmonary parenchyma. Cases of gastric cancer with 
involvement of portal, lumbar, and inguinal nodes, v. Recklinghausen referred 
to retrograde passage of obstructed lymphatics. Metastases in supra- 
clavicular nodes from tumors of abdomen and testis, with invasion and 
occlusion of thoracic duct have been explained by several observers as retro- 
grade lymphatic growth (Troisier, Poncet, Most). A cancer of gall-bladder 
with retrograde metastases in renal lymphatics, the veins being free, has been 
described by Vogel. With cancer of pancreas is sometimes found continuous 
lymphatic invasion down the aorta to the pelvic nodes and discrete nodules 
along the mesentery. In many of these cases it may be difficult to eliminate 
backward permeation, as Handley has shown, and some writers, especially 
Ribbert, doubt or deny the occurrence of retrograde lymphatic embolism. 
Yet there seems to be good Teason for holding that such embolism frequently 
occurs especially when a lymphatic trunk is occluded. 

That tissue and tumor-cells pass through the pulmonary capillaries 
has frequently been demonstrated and in certain tumors, especially the 
lymphosarcomas, general metastases develop in this way. Passing the lungs 
the emboHc cells tend to lodge in organs with feeble circulation, as liver, 
bone-marrow, and subcutaneous tissue. Or being arrested in the lung the 
embolic cells may grow through the vessels and break into the pulmonary 
veins {secondary embolism). Zahn speaks even of a tertiary embolism, 
metastatic tumors of the liver producing pulmonary emboli which grow 
through the lung. 

Abnormal communications between the cardiac chambers may permit 
the passage of larger emboli into the aorta. In a case of sarcoma of the 
seminal vesicles and in one of thyroid cancer, metastases were found in the 
kidneys but not in the lungs, from which Zahn concluded that emboli had 
passed through a patent foramen ovale, a process which he termed paradoxic 
embolism. Finally the blood-vessels are reached by way of the thoracic 

Thus invasion of the blood-stream may occur by: (i) Direct invasion of 
vessels in primary or secondary tumors, (2) secondary embolism, (3) tertiary 
emboHsm, (4) retrograde transport, (5) paradoxic emboHsm, (6) passage 
of single cells through pulmonary venules, (7) through the thoracic duct. 

Metastasis of a benign tumor was first claimed by Cohnheim for thyroid 
adenoma, and similar observations have since been reported by many 
writers. The great vascularity of this gland, the immediate contact of 
epithelium with capillary endothehum, and the extreme hyperplasia which 
may follow functional stimuli, offer unusual conditions in the thyroid favor- 
ing the dislodgment of adenomatous alveoli and even of simple hyperplastic 
cells. The occurrence of such metastases has been accepted by the majority 
of observers (Borst), but Williams and many others believe that in all such 
cases a small malignant tumor exists in the thyroid which gives origin to the 
metastases. The question is further compHcated by the great difficulty 
-of estimating from their morphology the growth capacities of thyroid cells. 


It is a fact, however, that thyroid cells have unusual powers of proUferation 
and growth, and it would seem necessary to admit that under rare conditions 
thyroid adenoma and hyperplastic thyroid cells may give rise to distant 
benign tumors. Definite proof of such an event requires minute search of 
the main and any accessory thyroids, which has been lacking in my own and 
many other reports. 

Similarly for the benign hydatid mole and even for the normal placenta 
there is a very suggestive group of facts which has led Marchand and others 
to believe in the occurrence of mahgnant metastases of normal chorionic 

Pure chondroma not infrequently enters veins and its ameboid cells 
produce distant metastases. In the notable case of Ernst's a chondroma 
of the vertebral column advanced through the lumbar, adrenal, spermatic, 
azygos, and cava inferior veins into heart and lungs. Discontinuous metas- 
tases usually in the lungs have often been observed from chondroma of pelvis, 
scapula, fibula, etc. (Lit. Ernst). The metaplasia of undifferentiated tumor- 
cells derived from teratomas or sarcomas must account for some chondro- 
matous metastases from tumors of testis and probably from other organs. 

Metastases have been observed with leiomyoma of the uterus (Klebs, 
Langerhans, Beetson, Schlagenhaufer) ; of the stomach (Hansemann, Moser, 
Borrmann); and jejunum (Ribbert). In some cases the structure of the 
tumor has not varied from the usual benign myoma, but in others the cells 
were larger and nuclei large and oval. Some were cystic and vascular. 
Birch-Hirschfeld . speaks of the invasion of uterine lymph-vessels by 
myoma. Williams denies that the simple uterine myoma ever produces 

Contact Infection. — Contact of a cancerous surface with an opposed, 
abraded or inflamed mucous surface is a rare source of transfer of tumors. 
The best known example is the infectious lymphosarcoma of dogs which is 
commonly transferred by coitus. This and other tumors of lower animals 
may sometimes be transferred by rubbing a granulating surface with a portion 
of tumor. 

The cases of cancer a deux in man have figured prominently at times but 
in none has the actual transfer by coitus been proven. Demarquay collected 
134 cases of cancer of penis in only one of which transfer by coitus appeared 
possible. Of 27 such cases collected by Gueillott none seems to me beyond 
question. Yet the genuineness of some of these cases cannot be denied. 

Transfer of epithelioma of lower lip to upper is difficult to establish but 
seems to have occurred in a few cases (v. Bergmann). In a case reported by 
Hartmann and Lecene a cylindrical cell cancer of the cervix seems to have 
become implanted on the squamous epithelial surface of the vagina. 

Implantation of epithelioma of esophagus in the stomach has been 
widely accepted as a fact (Borst), but Borrmann has found continuous per- 
meation of the lymphatics connecting such tumors. Similar evidence has 
not been adduced against implantation of renal tumor fragments in the ureter 
(Butlin). Likewise the contact transfer of carcinoma has been reported 
from one point to another in the urinary bladder, vagina, vulva (Thorn, 
Walter), cheek (Lucke), thighs (Wilhams), stomach (Klebs), vocal cords 
(Semon), and from hand to conjunctiva (Kaufmann). In most of these 
cases it is difficult to eliminate lymphatic extension and primary multiple 
tumors so that the doctrine of contact infection has always excited a legiti- 
mate scepticism. Bucher, Kaufmann, and others have fully stated the 
many grounds for this scepticism and it is obvious that the evidence must be 
very complete before any case can be accepted. The transferred tumor 


should be of different structure from that arising in the implanted surface 
and the lymphatics should be free from invasion. 

"Genius loci," or the particular susceptibility of a tissue to develop 
secondary tumors is an interesting phase of the study of metastases. _ Fa- 
mihar examples are the predominance of pulmonary metastases in chorioma, 
which is simply explained by the intravascular position of this tumor; the 
bone-marrow metastases of hypernephroma and cancer of prostate; and the 
multiple cutaneous tumors secondary to lymphosarcoma testis. The rnech- 
anism of the circulation will doubtless explain most of these peculiarities, 
for there is as yet no evidence that any one parenchymatous organ is more 
adapted than others to the growth of embohc tumor-cells. The spleen, how- 
ever, seems to escape with peculiar frequency. 

Virchow observed that organs frequently originating cancers are rarely 
the seat of metastatic cancer. Yet Zahn collected i6 metastatic cancers of 
the stomach, and 26 invasions of the ovary in 366 cases of mammary cancer. 

The preservation of structural type in metastatic tumors is often remark- 
able, yet as a rule there is increased anaplasia and more rapid growth in 
metastatic tumors and especially in recurrences after operation. 

Adenocarcinoma of the alimentary tract invades the lymph-nodes and 
the peritoneum, and produces distant metastases while rigidly maintaining 
its large alveolar type. Adenocarcinoma of the stomach has been recognized 
in a metastasis over the scapula. The metastases of benign tumors are 
always typical in structure. The bone metastases of hypernephroma are 
usually recognizable. Metastatic melanoma is usually but not always 
pigmented. On the other hand the general metastases of chorioma tend to be 
overgrown by Langhans cells with loss of syncytium. Squamous epithelioma 
may yield spindle-cell tumors in the liver. 

In general there is a gradual loss of structure in metastatic tumors. 
The liver metastases of mammary cancer, hypernephroma, and chorioma 
may be indistinguishable. The more anaplastic the original tumor the less 
is the room for change in its metastases. Mixed tumors tend therefore to 
become simple and certain elements may be eliminated in metastases and 
recurrences. This rule is best illustrated in the great variety of simple 
metastatic growths arising from teratoma testis. A peculiar form of elimi- 
nation of epithelium is seen in the regional metastases of mixed tumors of 
the parotid which are often purely myxomatous or chondromatous. 

Tridermal metastases which rarely arise from teratoma of testis or ovary 
indicate either that all three tissues are contained in the cell embolus, or 
that undifferentiated cells unfold their potencies after lodging in distant 

The osteoplastic carcinoma of the prostate offers a remarkable change of 
type and tendency in its pulmonary metastases. Here osteosarcoma ap- 
pears to accompany the carcinomatous elements. A satisfactory explanation 
of this process cannot be given but Schmorl supposes that the original 
bone metastases incite osteosarcomatous growth from which embolic cells 
yield osteoid tumors in the lung. 

It is characteristic of certain cases of several tumors that a very small 
original focus gives rise to very bulky secondary growths. This course is 
followed by some cancers of the breast with large tumors in the axilla, by 
melanoma of the skin or cancer of gall-bladder invading the liver, by mucoid 
cancer of colon involving the peritoneum. On the other hand squamous 
epithehoma rarely produces bulky metastases. 

Recurrence. — After apparently complete removal of malignant and of 
some otherwise benign tumors the growth may recur. The recurrence is 


usually in the operation wound or its immediate neighborhood (local re- 
currence), or in adjacent lymph-nodes (regional recurrence), or it may appear 
in distant organs from metastatic emboli. Sarcoma often recurs at a distance 
from the original growth while local and lymphatic recurrence is the rule 
with carcinoma. 

Recurrence by implantation from abdominal tumors in the laparotomy 
wound or in the peritoneum is a well recognized hazard of operations. The 
return of many malignant tumors is prompt and local but many years may 
elapse between the removal of choroidal melanoma and its reappearance in 
widely distant regions. 

The very wide variations in the periods of recurrences are the result of 
many complex factors so that statistical reports in this field are of limited 
significance. Yet Williams has gathered a very interesting collection of 
such records. Among the most striking of these reports are the long inter- 
vals sometimes elapsing between operation and recurrence, in cancer of 
breast, 30 years (Heurteaux, Verneuil) ; rectum, 21 years (English) ; of tongue, 
18 years (Guinard); choroidal sarcoma, 14 years (Fischer and Box); uterine 
cancer, 15 years (Pomard). The records of repeated recurrence are also of 
interest as encouraging the surgeon. Thus Gross performed 22 operations 
in 54 recurrent tumors in a case of sarcoma of the breast during four years, 
the patient finally recovering. 

Recurrent tumors may reproduce exactly the structure of the original 
or follow the rules already stated for metastatic growths. Local recurrences 
usually show increasing malignancy and anaplasia, and at times the recurring 
tumor shows an astonishing variation from the original type. Clinical 
observation is here supported by the results obtained with transplantable 
tumors which by repeated passage may be considerably heightened in activity. 
Squamous epithelioma is comparatively stable in type, as is also adenoma 
destruens. Most alveolar carcinomas tend to lose the alveolar form and 
recur diffusely. Myxoma of the skin recurs repeatedly with the same struc- 
ture, while melanoma may retain or lose its pigment. Recurring mixed 
tumors tend rapidly to become simple. One sees pure chondroma in the 
returns of parotid tumors. In the second to fourth recurrences of a complex 
tridermal teratoma testis I found pure angiosarcoma. In a case of ada- 
mantinoma originating as squamous epithelioma, L'Esperance, in this 
laboratory, has recorded progressive changes through plexiform epithelioma, 
glandular carcinoma, diffuse small spindle-cell structures, alveolar round- 
cell sarcoma, and finally in the sixth operation, diffuse round-cell "sarcoma." 

There appear to be no indications that a recurrent tumor may be less 
anaplastic than the original, but Apolant claims to have reduced an alveolar 
carcinoma to an adenoma by transplantations in immunized mice. 

Recurrent tumors in the great majority of cases arise from cells or portions 
of the tumor which have escaped removal. Microscopic section of the edges 
of specimens removed at operation often reveals that the knife has passed 
through strands of tumor tissue or vessels containing tumor-cells. There 
can be little doubt that the rough manipulation of cellular tumors in the 
preparation of the patient and in the excision of the growth widens the field 
of infection by forcing cells through vessels and tissues spaces. Or, viable 
cells are scattered through the wound by piece-meal removal from inaccessi- 
ble regions. The practice of removing a portion of the tumor for diagnosis 
may add to the dangers of local dissemination. With myxoma, a tumor 
which is characterized by persistent local recurrence, the semi-fluid tissue 
may very readily be forced into the surrounding fat tissue. 

Arising from remnants of the original growth recurrence is usually 


prompt, multiple, in the line of incision or near by, and of the same type as 
the original growth or more maUgnant. Yet there are many records of cases 
in which recurrence in the line of incision has been delayed for years. Such 
cases require the conclusion that displaced cells may long remain dormant. 
The many instances in which partial removal of malignant tumors has been 
followed by regression of the remnant, as in ovarian carcinoma and in 
chorioma, indicate that the mechanical disturbance of nutrition, or removal 
of local sources of growth stimulus, or the general reHef of the organism 
from toxic products, turn the balance in favor of the patient, and lead to 
absorption of tumor-cells. Thus tumors in rats sometimes regress after 
hemorrhagic infiltration following crushing trauma. The same factors may 
be assumed to retard the growth in more malignant tumors which spring up 
at varying periods after incomplete removal. 

As a second source of recurrent tumors it has been suggested by v. 
Rindfleisch, Beneke, and others that tumors exert a pervasive formative 
influence leaving the surrounding tissues in a state of excitation and with a 
momentum toward neoplastic growth, so that after the removal of the tumor 
the adjacent tissues give origin to a new tumor. This somewhat vague 
conception is vigorously rejected by Ribbert and is quite irreconcilable with 
the theory of origin from isolated cell groups. Yet there are several considera- 
tions which render it inadvisable to discard such a theory of recurrences. 
This question is closely related to that of the influence of tumor growth upon 
surrounding tissues and to that of the multiple origin of tumors in the same 

Attention has already been called to the fact that some tumors advance 
by gradual lateral extension over previously normal cells, so that removal 
at one stage of this extension may leave other cells which later reveal their 
momentum toward tumor growth. The local recurrences of some lingual 
and laryngeal cancers seem to accord with this view. Further, the local 
predisposition to tumor growth which figures distinctly in many phases may, 
as Borst points out, extend over a considerable area in all of which the tumor 
growth does not reveal itself at once but develops in separate stages or foci. 
Such conditions seem rather clearly revealed in neurofibromas, and, according 
to Hauser, with various carcinomas, especially those of the gastro-intestinal 

Ovarian cystomas recur in the opposite ovary or in fragments of the 
first ovary or in misplaced ovarian structures in the neighborhood, showing 
the influence of local predisposition (Velits). Many authors have pointed 
out the great frequency of cancer after ovariotomy and have endeavored to 
show, without 1 think complete success, that loss of the ovaries greatly 
increases the local disposition of many tissues to cancer (Wells, Kratzenstein, 
Neugebauer, WilUams). 

A third source of recurrence is found in the multiple origin of many tumors. 
The removal of a small cancerous nodule from the breast with chronic mastitis 
may readily be followed by the appearance of a new tumor in an adjacent 
portion of the gland or in a supernumerary gland. One seldom hears of 
recurring uterine myomas since these are so clearly multiple tumors, but 
multiplicity while less obvious is perhaps equally common with epithelioma 
and sarcoma of the skin where the new tumor is often regarded as a recurrence. 

New tumors following operation differ from the true recurrences in not 
appearing in the scar, developing usually after an interval, reproducing the 
original growth without great change in maUgnancy, and often being solitary. 


Constitution of Tumor Proteins. — The conception that tumor proteins 
must differ in essential respects from those of normal tissues has not been 
demonstrated by chemical methods. The nature of the problem involved 
appears to have presented itself in different forms in the minds of investi- 
gators who have attacked the problem. 

A different distribution of normal proteins from that of normal tissues 
has been demonstrated in tumor tissue by Petry, Wolff, and Beebe, who found 
a higher content in nucleoprotein, more uncoagulable protein, and less 
globulin and albumin. It is probable that these results depend on the over- 
growth of cell nuclei, degenerative and autolytic processes, and edema. 
Nucleohiston is present only in lymph-nodes among normal tissues and its 
presence in lymphatic metastases ot tumors originally free from this protein 
indicates that metastatic tumors receive chemical impress from the tissue 
in which they are growing. Although nucleohiston is absent in primary 
carcinoma of breast and in the testis, Beebe found this substance in lymph- 
node metastases of mammary cancer and Bang in the lymphatic metastases 
of testicular carcinoma. Nucleohiston gives a precipitate when CaCl2 is 
added to a watery solution of the tumor. 

Direct chemical analj^sis of the split products of cancer proteins by Wolff 
yielded a high proportion (35 per cent.) of glutaminic acid, while Bergell 
and Dorpinghaus found excess of alanin, phenylalanin, asparaginic acid and 
diamine acids. Yet these results conflicted with those of Petry, Neuberg, 
and Beebe. 

Resistance to peptic and susceptibility to tryptic digestion was said by 
Blumenthal and Wolff to distinguish tumor from normal tissues. Yet 
their results were not uniform and were probably determined by the increased 
amount of nucleoproteid in some of the tumors. 

Excess of potassium and deficiency in calcium in rapidly growing tumors 
free from necrosis, and the opposite relations in slowly growing or old and 
necrotic tumors have been demonstrated by Beebe and by Clowes. 

Pentose was greatly increased in a fibrocarcinoma of the breast in com- 
parison with the amount in the normal breast in cases studied by Beebe 
and Shaffer. These authors also found that the pentose content in different 
tumors varied and bore no relation to the nucleoprotein, or to the presence 
of degeneration. In hepatic carcinoma and in hepatic metastases of gastric 
carcinoma Neuberg found 70-80 per cent, more pentose than in normal liver. 

Lactic acid appears in tumors, according to Fulci, in considerable 
quantities. It is more abundant in epithelial than in connective tissue 
growths and increases in the more malignant actively growing tumors. Its 
formation is dependent on metabolic activity of the tumor-cells, its source 
the blood carbohydrates and possibly the proteins. It appears to have no 
relation to cachexia. 

Of the total phosphorus of the normal liver B. Wolter found 24.68 per 
For a fuller discussion of this subject with literature consult Wells, "Chemical 



cent, as phosphatid phosphorus, in the tumor-free portions of a Uver with 
primary carcinoma 22.04 per cent., and in the tumor nodules 16.28 per cent., 
while the protein phosphorus in the same materials ran 20 per cent., 25.5 
per cent, and 26.70 per cent. In 0.0634 gm. of dried tumor substance he 
found 1.40 per cent, of cholesterin. 

Tryptophan was markedly increased in an epidermoid carcinoma of skin, 
and in an hepatic carcinoma, over the proportions found in normal skin and 
Hver in cases studied by Fasal, but this substance was absent in a mammary 

Fats. — The chemistry of tumor fats has been studied extensively in renal 
and adrenal tumors {q.v.). In general it appears from Bossart's work that 
actively growing tumors free from degeneration contain little fat and much 
lecithin, while with degeneration and necrosis free fats replace lecithin. 

Purin bodies were found by Wells and Long in about the same form 
and amount as in normal tissue and less abundantly than the nuclear 
content would suggest. The purin enzymes were also identical with those of 
normal tissues, guanase being constantly present and adenase absent. 

The delicate methods of immunological studies indicate that there 
are biological differences between certain tumor-tissues which are probably 
based on chemical distinctions but the results obtained in this field are not 
decisive. Following Michaelis' failure to produce specific immune bodies 
against mouse tumor cells, Beebe working with purified nucleoproteids of a 
leukemic spleen produced a serum which agglutinated the emulsified cells of 
this spleen and those of a lymphosarcoma, but acted feebly and only in strong 
concentration on cells and nucleoproteids of normal spleen and other tissues, 
as well as of cancer and spindle-cell sarcoma. 

Other data bearing on this question are considered in the discussion of 
experimental immunization. 

Tumor Ferments. — The study of special ferments in tumors by Buxton 
and Shaffer demonstrated no distinct differences in quantity or quality from 
equivalent normal tissues. Weil using the viscosimeter found more proteo- 
lytic activity in certain cellular tumors than in certain normal tissues but was 
not prepared to say that the difference did not depend on leukocytes. Abder- 
halden, however, concluded that extracts of tumor-tissue and those of normal 
tissue split certain polypeptids in a different manner. Extracts of normal 
mouse liver cleave certain polypeptids slowly while those of mouse tumor act 
much more quickly. In comparing the extracts of mouse and human 
tumors with those of normal tissues he found differences in the split products 
of the proteins after the action of these ferments. Comparing the pepto- 
lytic action of normal and tumor-tissue from lower animals on peptone, 
Abderhalden and Medigreceanu found occasional but no constant or striking 

Autolysis is often observed to proceed more rapidly in certain tumor 
tissues than in normal tissues, but it is probable that all such differences 
depend on the more cellular character, and presence of degenerating tissue, 
edema, leukocytes, and bacteria. It is extremely difficult to obtain normal 
tissue which may safely be compared with tumor-tissue in this respect. 

The increased activity of autolysis is well illustrated in Yoshimoto's 
experiments in which an hepatic carcinoma yielded 7.2 gms. N in the split 
products as compared to 4.8 gr. in equivalent units of normal liver. With 
a mammary carcinoma the difference was even greater. In the tumor auto- 
lysate purin N was reduced, while that of diaminoacids, peptone,; and 
ammonia was increased in proportion. 

It is commonly believed, on the basis of Jacoby's experiments, that 


autolytic ferments have a high degree of specificity for the proteins of the 
organs in which the ferments are found. Blumenthal and Wolff have reported 
that when measured amounts of tumor-tissue and of normal tissue are auto- 
lyzed separately in one series, and conjointly in another the autolysis is 
always greater in the conjoined series. They conclude that the tumor ferments 
attack the normal tissues and exhibit a heterolytic property. On this basis 
rests the claim that infiltrative growth and cachexia depend on the heterolytic 
activities of the tumor ferments. Baer and Ettinger demonstrated a pro- 
teolytic activity in cancerous ascitic fluid which failed to appear with 
ascitic fluid from other sources, but Kepinow and Hess and Saxl were unable 
to verify any of these observations. The possible presence of bacteria, 
necrosis, or postmortem decomposition has not apparently been considered, 
although Neuberg has subsequently denied that bacteria or leukocytes have 
any influence on the results. 

Ascitic fluid of carcinomatous origin yielded little euglobulin (12.44 
per cent.), increased albumin (64.92 per cent.), and the usual proportion of 
total globulin 35.07 per cent. (Joachim, Wolff). Similar proportions of 
proteins were determined in edema fluid and expressed cancer juice by Wolff. 

The milky character of carcinomatous ascitic fluid is ascribed by Wolff 
to the presence of cholesterin-acid-ester combined with euglobulin. In 
a series of cancerous and other ascitic fluids Weil demonstrated occasional 
but inconstant hemolytic and hemagglutinative properties, abundance of 
complement in two cancer cases, and some antitryptic action, but no constant 
distinction between the cancerous and non-cancerous fluids. Signs of active 
autolysis in a peritoneal exudate accompanying ovarian carcinoma are 
reported by Umber, who found increase in non-coagulable N, albumose, 
leucin and tyrosin. Eppinger reports similar findings. In a bloody cancer- 
ous exudate K. Wiener demonstrated an ereptic ferment, but any tryptic 
ferment present was masked by the blood. In this fluid were traces of 
histidin and arginin. 

J. W. Vaughan has observed a striking relation between anaphylactic 
sensibiHty of guinea-pigs inoculated with tumor residue and the lymphocytosis 
excited by the inoculation. Animals receiving the water-soluble residue of 
cancer-cells after their extraction with alkahzed alcohol, or a vaccine of 
tumor-cell emulsion after extraction with alcohol, showed in many cases very 
high lymphocytosis, and in such sensitized animals anaphylactic death was 
readily produced. The sensitization was quite transitory, lasting only during 
the 4 to lo-hour period of lymphocytosis, and the author interprets the 
result as depending on the action of lymphocytic ferments on the tumor-cehs. 

Wassermann Reaction in Cancer. — The results obtained with the Wasser- 
mann reaction in cancer patients have varied very widely. Caan secured 
positive reactions in 41 per cent, of 85 cases; in 6 of 7 lip carcinomas; in 
9 per cent, of breast tumors; and in 17 per cent, of gastro-enteric tumors. 
All of these cases were free from chnical signs of lues. A. Foerster secured 
36 negative reactions in 37 cases, and Noguchi 38 in 39, while F. J. Fox 
reports 5 positive reactions in 210 cases of cancer. Positive reactions in 
patients with cerebral tumors free from syphilis appear to be not uncommon 

Complement Deviation with Autogenous Tumor Antigen.— In over 100 
cases of carcinoma and 16 cases of sarcoma the complement deviation test 
of the blood serum has been made, using tumor antigens of the same or 
similar tumors. In the majority of cases the test has been positive, but in a 
large number the results have been negative. Recently, W. Barratt, using 
the patient's own tumor as antigen, got negative results in six cases. De 


Marchis often obtained negative results in both early and advanced cases. 
A positive result he regarded as of considerable diagnostic value although it 
occasionally occurred in syphiHs and other diseases. With an antigen pre- 
pared from human carcinoma v. Dungern claims to have obtained comple- 
ment deviation with the serum of both carcinoma and many other tumors 
such as lymphosarcoma, glioma, and even myoma. These paradoxical 
i-esults failed of verification in the hands of Edzard, Schenk, and Coca. 

It is evident that the subject of complement deviation in tumors is without 
substantial foundation. 

Complement deviation by antigen of Micrococcus neoformans tested 
with serum of cancer patients, in all but 7 of 144 cases, was secured by Yama- 
nouchi and Lytchkowsky. The reaction was also positive in 44 cases of lues, 
but not in benign tumors. F. Green obtained positive reactions in cancer 
with the antigen of Micrococcus neoformans and also with those of staphy- 
lococci and streptococci. 

Antitryptic. Power of the Blood. — Blood serum has long been known to 
inhibit the action of certain ferments. Brieger and Trebing first found that 
the serum of cancer patients inhibits very markedly the action of trypsin. 
Using various technical methods, these authors, followed by Bergman and 
Meyer, Herzfeld, Roche and others, showed that in about 90 per cent, of 
cancers there was marked increase in the antitryptic power of the blood serum 
but that considerable increase occurred also in many other diseases, especially 
in those attended with leukocytosis (Wiens, Schlecht). Weil by an exact 
quantitative method determined that the "antitrypsin" is increased in some 
cancer cases beyond that observed in any other diseases, while in other cases 
it failed to fall as low as the ratio observed in other diseases. All authors 
agree that the failure of any increase is rather strong evidence against the 
existence of established malignant disease. Brieger regarded the phenome- 
non as a sign of cachexia and an immunity reaction against excessive amounts 
of proteolytic ferments derived from tumor-cells. Wiens and Schlecht 
trace a close relation with leukocytosis. Weil points out that the exact 
nature of the inhibiting substance is undetermined and that no specific relation 
to trypsin need necessarily exist, since blood serum also inhibits saponin. 
Among 57 cases of cancer and three of sarcoma S. M.Lewin found much in- 
creased antitryptic power in 57. He could not refer the antiferment action 
to leukocytes but attributed it to a reaction to proteolytic ferments discharged 
into the blood from degenerating tumor-tissue. 

Meiostagmin Reaction. — This reaction is based upon a change in the 
surface tension of blood serum. When to a dilute solution of an alcohol- 
ether extract of cancer tissue is added a small portion of blood serum from a 
carcinoma patient, a change in the surface tension is produced which may 
be estimated by the reduced number of drops of the fluid passing in a 
given time through a specially designed pipet or stalagmometer. This 
physiochemical alteration was employed by Ascoli in the study of cancer 
serum and was considered by him to result from a somewhat specific inter- 
action of cancer antigen and cancer antibody developed in the serum. The 
specific character and immunological nature of the reaction was soon set 
aside when it was discovered that extracts from many organs and even syn- 
thetic substances, as linoleic and ricinic acids, were equally effective as antigens 
(Ascoli, Izar). It was also shown that the sera of a great variety of diseases 
and of pregnancy reacted in the same manner with a variety of antigens, 
thus establishing that the production of the reacting factors was the result 
of a constitutional process of widespread occurrence. 

To carry out the test one may best employ a solution of 0.2 gm. ricinic 


acidpluso.sgm. of linoleicacidin lo c.c. absolute alcohol. To i ex. of serum 
is added, by dropping carefully from a pipet, o.oi to 0.02 c.c. of the antigen 
solution and then 9 c.c. of 0.85 per cent. NaCl with thorough mixing. As 
a control i c.c. of serum is added to 9 c.c. of salt solution. Both tubes 
are kept i hour on a water bath at 5o°C. and after 2 to 6 hours the drops 
are counted by the stalagmometer. The count increases in the sera receiving 
antigen, an increase of i.,^ or less being regarded as negative, 1.7 + posi- 
tive, intermediates uncertain. 

Various other antigens may be employed according to the choice of the 
investigator. Burmeister prepared several antigens by different methods 
all with practically identical results. The results obtained vary considerably. 
Ascoli and Izar secured nearly constant figures with tumor sera. N. Blu- 
menthal (1915) finds that 78 per cent, of gastro-intestinal carcinomas react 
positively, while in other forms the positive results are much less, and in 
cutaneous and hepatic carcinoma and sarcomas very few. He concludes 
that the test is of diagnostic value only in gastro-intestinal carcinoma, and 
in pregnancy, while it is always necessary to exclude diabetes, uremia, in- 
fectious fever, tuberculosis, lues, chronic joint diseases, and cirrhosis. The 
diagnostic importance is therefore practically nil. 

Various hypotheses have been suggested in explanation of the reaction. 

1. The presence in the serum of abnormal lipoids resulting from fermen- 
tative processes in the cancer-cells. This view is favored by the heat-resist- 
ance of the serum; by the ready removal of the active factors by extracting 
the serum with alcohol, ether or chloroform (Izar, Michaeli, Cottoretti) ; 
by the discovery of synthetic antigens as linoleic acid (Kohler, Luger); 
and by the demonstrated increase of lipoids in reacting sera (Michaeli, 
Cottoretti). The reaction may also depend on a diminution in blood, 
cholesterin (Izar). 

2. A diminution in blood albumens, and presence of peptone. This 
condition is observed in most diseases giving the reaction. 

3. Changes in the alkalescence of the blood. Heating the serum in- 
creases the reaction and drives off CO2 from the bicarbonates. 

The epiphanin reaction of Weichardt depends upon a change in the alka- 
lescence of the serum. This reaction has been found positive in 81 per cent. 
of a series of cases studied by Jozca and Tokioka. By this means E. Rosen- 
thal finds it possible to demonstrate antibodies in the blood of guinea-pigs 
immunized by human or mouse tumor and in the serum of tumor-bearing 

Stammler's Reaction. — Most tumor extracts exhibit a slight or distinct 
opalescence. Stammler observed that the addition of cancer serum clears 
up this opalescence with the formation of a slight precipitate, while most 
normal sera fail to act in this manner. 

Ransohoff's Test. — Ransohoff observed that the injection of 3-5 c.c. of 
blood serum from cancer patients in guinea-pigs which had previously been 
sensitized by i c.c. of such sera, caused none or very weak anaphylactic 
reaction, while normal human serum produced the usual violent reaction. 
Since the same immunity followed sensitization by tumor extract, he con- 
cluded that the immunizing substance of the serum was derived from the 
tumor. In a series of 30 cancer cases 92 per cent, gave a positive reaction 
while the test was always negative in other conditions. F. Green was 
unable to verify these results. 


The Embryonal Theory. — For many years before the appearance of 
Cohnheim's work, observations had been accumulating to show that tumors 
were in some way related to the embryonal growth of tissue. Lobstein, 
1829, likened the growth of a tumor to that of embryonal tissues and con- 
ceived that neoplastic growth had lost the control of the organism. Recamier, 
1829, noted that cancer developed from irritated moles and pointed out that 
supernumerary organs, " tous les tissues extraordinaires accidentals," readily 
degenerated into cancer. Rokitansky regarded certain myxomas, in his 
class of collonema, as derived from embryonal connective tissue. Houel, 
1864, designated certain sarcomas as embryonal, because they seemed to 
represent an abnormal growth of the same elements which in the embryo 
formed normal organs and tissues. In the subsequent decade there were 
several observations of tumors arising in connection with abnormalities of 
development, and the embryonal theory was believed to apply to the group 
of complex teratomas (Lucke). A further basis of the embryonal theory 
existed in the belief maintained by Remak, 1854, that cancer arose from 
misplaced islands of epithelial cells in tissue not normally containing epithe- 
lium. Paget in 1853 remarked that invisible^ defects in the formation of 
organs might render them or portions of them peculiarly apt for the seats of 
malignant disease. 

In 1874 Durante clearly stated the doctrine that all tumors arise from 
embryonal groups of cells. He observed sarcoma twice developing from 
pigmented moles, examined many nevi, and described the embryonal character 
of the cells composing them. Hence he concluded that similar circumstances 
must surround the origin of all tumors, especially the malignant growths. 

The modern embryonal theory was placed on a comprehensive basis by 
Cohnheim whose original conceptions were ably supported by his own ob- 
servations and have been steadily strengthened by very numerous contribu- 
tions from many sources during the past forty years. While not a theory 
of universal application, that embryonal cells possess more than any others 
the essential factors of tumor growth is, perhaps, the most important single 
fact in our knowledge of tumor genesis. 

Cohnheim believed that tumors develop from masses of simple or complex 
tissue misplaced during embryonal development. Or, they arise from small 
groups of superfluous cells which have retained their embryonal characters 
but are not necessarily misplaced. The idea of the embryonal character 
of the cells appeared to him essential. Most of the embryonal cells result 
from overproduction after the formation of the germ layers and before the 
appearance of definite rudiments of the organs, and he held that these super- 
fluous cells were either distributed throughout -the viscera or were gathered 
at certain points, as the mucocutaneous junctions. The sudden develop- 
ment of the cells he referred chiefly to changes in the blood-supply. 

The entire group of mixed tumors and the simple heterologous growths 
were at once included in the scope of the embryonal theory. Indeed when 
one explores the group of teratomas there seem to be all intermediate forms 



from the complex sacral tumors up to parasitic implantations containing 
parts of organs or limbs and even up to such phenomena as identical twins. 
Likewise when one passes downward through the complex heterologous 
tumors of the parotid to the simple heterologous growths which arise from 
small groups of cells misplaced or remaining embryonal it would seem that 
tumor genesis were chiefly a question of the mechanics of development. 
The present support of Cohnheim's theory is extensive. Congenital tumors 
of many varieties, chiefly mesoblastic but also epithelial, exist at birth or 
develop shortly after, and the evidence is clear that embryological disturb- 
ances are concerned in their origin. Such tumors include fibroma, myxoma, 
lipoma, chondroma, angioma, gHoma, myoma, mixed tumors of kidney, 
dermoids, and teratomas. 

Congenital cancers affect the kidneys (Leibert), adrenal (Reiman), 
testis (Phillip), vagina, ovary (Ahlfeld), pylorus (CuUing worth), pancreas 
(Bohn),skin (Braun), shoulder (Selberg),neck (Kronlein),pleura(Muus),etc. 

The embryonal tumors of adults which yield a type of tissue closely 
simulating the embryonal counterparts and resemble the congenital tumors 
of the same regions form a well-recognized group for which an identical 
etiology seems highly probable (Wilms) . Among such tumors are myxomas, 
rhabdomyoma of heart, amyelinic neuroma, tubular neuro-epithelioma, and 
a great variety of very cellular malignant tumors whose histogenesis has not 
been fully traced. Rudimentary organs are specially susceptible to tumor 
growth, as undescended testes, Luschka's gland (Defosses), the paro-ophoron 
(Ricker), the neurenteric canal (Kraske). 

Supernumerary organs are frequent seats of tumors. The misplaced 
adrenal rests have a wide distribution and give origin to many heterotopic 
growths, in and about the kidney, along the genito-urinary tract, in testis 
and ovary, throughout the pelvis, and elsewhere (Chiari, Askanazy, 
Lubarsch). Supernumerary breasts are occasional sources of mammary 
cancer. Aberrant portions of thyroid, thymus, pancreas, uterus, and ovary, 
give rise to many tumors in the regions of these organs. Some authors 
assume the presence of straggling sex cells all the way from the cephalic 
to the caudal extremities of the embryo to account for complex teratomas in 
neck, thorax, and abdomen. 

The association of tumors with local abnormalities of development 
strongly supports the embryonal theory. Prominent in this category are 
the melanomas arising from pigmented moles and pigment spots in sclera 
and choroid; the gliomas associated with spinal hydromyeha and various 
abnormalities of cord and brain; sarcoma in cystic kidneys; adenomyoma of 
uterus; gliomas, dermoids, and mixed tumors, with spina bifida; adenoma 
with cysts of liver. 

Persistent remnants of embryonal structures clearly originate the bran- 
chiogenic cysts, epitheliomas and chondromas. The thyroglossal duct 
gives cystic or soHd tumors sometimes hned by ciliated epithehum at base of 
tongue, in floor of mouth, or deep in the neck. Similar cystic tumors with 
cihated epithelium occur also in esophagus, peritoneum, and liver. Certain 
embryonal tumors of uterus and adnexa and bladder are probably derived 
from remnants of Wolfi&an body or Gartner's duct. The pars postanalis 
intestini yields epithehal tumors of the lumbar region, and the omphalo- 
mesenteric duct and urachus give rise to abdominal tumors. 

The extensive lists of heterologous tumors are most readily explained as 
growths from aberrant cell groups. These include chondromas of tonsil, 
breast, parotid, bladder, uterus, thyroid; epithehal tumors in connective 
tissues and organs not normally containing such cells; the mixed tumors of 


uterus, bladder, breast; lipoma of pia and brain, and myo-lipoma of sper- 
matic cord (Sazarin). \ATiile the heterologous tumors and teratomas of the 
sex glands are probably derived from sex cells, the origin of teratoid tumors 
in abdomen, thorax, pharynx, and other organs may stiU be referred in part 
to aberrant cell groups of another t3rpe. 

The predilection of tumors for mucocutaneous junctions, ostia at the 
points of fusion of embryonal structures, and fissures formed by incomplete 
union of such structures, impressed Cohnheim and aU later observers as 
strongly favoring the theory of origin from superfluous cells. 

Here must be mentioned the frequent cancers of lip, nares, anus, rectum, 
cardia, pj-lorus, portio vaginaUs uteri, rete testis, and those at the tracheal 
and esophageal junction where squamous and cihated epithehal tumors have 
been found transposed across normal boundaries. Misplacements at em- 
bryonal fissures must be connected with the cholesteatomas and epidermoids 
of skin, breast, abdominal and cranial cavities, spinal canal, and middle ear. 
]\Iany chondromas of the skeleton, as multiple chondroma of spine and spinal 
canal, trachea, etc., are referred to irregularities in formation and separation 
of islands of cartilage. The chordomas of spine, base of skull, and nares 
reproduce the embryonal chorda dorsahs. Dentigerous cysts, odontomas, 
and the frequent adamantinoma of alveolar borders arise from superfluous 
or isolated remnants of the teeth. Symmetrical tumors sometimes strongly 
suggest embryonal irregularities, examples of which are symmetrical nevi, 
xanthomas (Rayer, Ehrmann), hpomas (Grosch, Kottnitz), angiomas, 
myomas (Brigidi, Marcacci), sarcoma of epiphyses. 

There are recorded many notable examples of symmetrical tumors in 
paired organs, as fibrohpoma of kidneys, adenoma and carcinoma of kidneys, 
ovaries, tubes, breast, adrenals, and lymphosarcoma of testis. Multiple 
systemic tumors of the nerve-trunks, neurofibroma, suggest an imperfect 
differentiation of endoneurium. From this point one is led into the exten- 
sive field of multiple primary tumors of the same or different organs. Studies 
in this field have usuall}^ led to the conclusion that developmental eccentrici- 
ties were involved. 

Later foUowers of Cohnheim's suggestions, especiaUy Ribbert and \\'ilms, 
have urged that in extra-uterine life cell groups may become isolated and 
superfluous by various mechanisms, and form the basis of tumors. The 
history of phylogenetic and ontogenetic development suggests that certain 
tumors signify spasmodic reversions to the anatomic conditions of prehis- 
toric man or of other closely related animal species. Thus the fusion of 
multiple renculi, the coalescence of multiple uteri and breasts, the reduction 
in length of stomach, intestine, appendix and colon, loss of lymphoid tissue 
in the relatively narrow human cecum, and the elimination of hair and 
sebaceous folhcles in the face, suggest a source of superfluous cells and lay a 
significant foundation for the atavistic growth of tumors in many organs. 
Such influences, however, probably form only the feeblest of the predisposing 
causes of tumors. 

The studies of R. Williams, Lubarsch, Ribbert, Borrmann, Meyer, and 
many others, have shown that embryonal rests are far more frequent than was 
at first imagined. Wiesel for example, finds that adrenal rests are almost 
constantl}^ present along the spermatic cord and in the pelvic tissues of both 
sexes. It is possible to argue with fair success, as R. Williams has done, that 
all known tumors of the uterus arise from aberrant, or superfluous, or embryo- 
nal cell groups. From a study of the various tumors of the testis I have 
been led to the conclusion that practically all of them are derived from sex 
cells. For tumors of the ovary Ribbert reached a similar conclusion. Thus 


the more careful analysis of the origin of tumors has extended rather than 
restricted the number of tumors that are known to arise from cells that have 
lagged behind in development. 

Limitations of Cohnheim's Theory. — The pursuit of the embryonal theory 
has added extensively to our knowledge of aberrant cell groups and at the 
same time revealed many new difficulties in the way of the embryonal theory. 

It has been shown that neither the isolation nor misplacement, nor 
abnormal persistence of cell groups, are necessarily followed by tumor growth. 
The fate of tissue rests varies: (i) they may remain stationary; (2) they 
may pass through normal stages of development and eventually atrophy; 
(3) they may experience only a limited growth; (4) cysts may form from 
them, or (5) tumor growth may appear. 

Various hypotheses have been suggested to account for neoplastic growth 
from rests. The period of isolation of the cells is an important factor. The 
earlier its occurrence the less is the differentiation and the greater the capacity 
for growth. Early embryonal rests when starting to grow meet conditions 
which do not favor normal development. 

Since only certain rests produce tumors it may be that the embryo- 
genie disturbance which produces the rest stamps the cells with abnormal 
qualities (Aschoff). This disturbance may perhaps be associated with 
failure of the cells to experience their proper idioplastic development and 
thus leave them in a condition unusually favorable to growth when such 
potencies -are in excess. Or the relation to the surrounding tissue may in- 
fluence the growth of the rest. Separation from the normal nervous control 
of the tissue is here to be considered. Organized cell groups such as those 
composed of epithelium with supporting connective tissue, according to 
Ribbert, tend to develop in an orderly manner, while epithelium alone 
tends to develop a malignant tumor. Considering the great number of 
organized adrenal and thyroid rests, tumors from these sources are rare. 
The attempts to produce mahgnant tumors experimentally by transplanta- 
tion of tissues, adult and embryonal, have not succeeded. On the contrary 
the very numerous studies in this field have served to show that something 
more than the separation of cell groups from their natural environment 
is necessary for progressive growth (cf. Section on Experimental Research). 

Thus it becomes evident that Cohnheim's theory while it explains the 
structure and occurrence of many tumors wholly fails to reveal why the 
embryonal cells begin to grow and when growing produce malignant tumors 
instead of normal structures. 

The limited growth of many benign tumors has been approached by 
the structures produced by transplanted tissues, but the complete eman- 
cipation of malignant neoplasms from the normal laws of growth remains 
the obscure and essential element. In the attempt to solve this difficulty 
the theory of cell autonomy has grown up. 

Theory of Cell Autonomy.— The defects of Cohnheim's theory were 
brought to light by studies undertaken in its defense. They showed that the 
mere presence of embryonal cells was not sufficient to account for their growth 
in tumors, and that tumors grow where probably no embryonal cells exist. _ It 
is necessary to consider how tumors arise from cells which are neither origi- 
nally misplaced nor essentially embryonal. In this field the theory of cell 
autonomy has developed. 

The germ of the theory of cell autonomy appeared in the studies of 

Remak and Thiersch who traced the antagonistic relation of epithelium and 

connective tisue throughout embryonal development. Epithehum seemed 

everywhere to be the dominating embryonal tissue and to cease growing when 



it met sufficient resistance from the connective tissues. Thiersch found 
in the weakening of the stroma of the involuting breast a rehef of tension 
capable of releasing the formative tendencies of gla;id-cells with their 
longer span of vital activity . The decay of connective tissues he regarded as 
senile atrophy and the growth of cells as degenerative prohferation. The 
idea of cell autonomy appears also in Cohnheim's theory but here it was over- 
shadowed by the embryonal element in the originating cells. 

To account for the Hmitless growth of cells, embryonal or adult, the 
theory of cell autonomy introduces the conception of tissue tension. 

Ribbert endeavored to explain the growth of tumors solely through 
the removal of certain cells from the influence of a tissue tension by whidi 
their growth is normally restrained. He abandoned the theory that the dis- 
turbed cells must be embryonal, claiming that the regenerative capacities of 
adult cells are quite sufficient to account for all the characters of malignant 

Conception of Tissue Tension. — The theory of tissue tension assumes that 
cells are capable of growing indefinitely unless restrained by a complex 
group of forces. This assumption is perhaps not yet proven but there are 
many facts in its favor. The astonishing regenerative powers of injured 
tissues in plants, lower animals, and of many organs and tissues of ver- 
tebrates, the remarkable rate of growth of the vertebrate embryo, the hy- 
pertrophy of the gravid uterus, the indefinite reparative capacity of injured 
or transplanted epidermis, all suggest that normal cells, although eventually 
restrained by organized forces, yet in proper environment are capable of 
growing without limit. Ribbert insists that no unusual power of prolifera- 
tion exists in cancer-cells, that these cells freed from the restraints of tissue 
tension are merely exhibiting the powers of growth with which they are en- 
dowed from the ovum. Weigert and Roux have also asserted that the re- 
generative capacities of cells are determined from the moment of their de- 
rivation from the ovum and can never be increased by any external stimulus. 

Many observers have not accepted the Weigert-Roux hypothesis. 
Lubarsch finds it incompatible with the extensive growth of metastatic tumors 
and claims that external stimuli may induce an enormous increase in the re- 
generative capacities of cells. I believe it is here necessary to distinguish 
between the organized regenerative properties of tissue cells, which are truly 
determined in the ovum, and the mere power of multiplication without or- 
ganization which is subject to immense variations from the influence of the 
environment. Tumor-ceUs seem to have more than their ancestral power of 
growth as when without pressure they erode bone. In any case the facts 
of normal and tumor growth seem to place the doctrine of growth restraints 
and tissue tension in an unassailable position. 

What, then, are the restraints to growth which control the multiplication 
of cells and maintain the tissue in a physiological condition while permitting 
normal repair? At least four such factors are known: 

1. Mechanical pressure of cells on each other. 

2. The distribution of nutriment. 

3. The influence of specialized functions. 

4. Organization. 

Under each of these heads belong important contributions to the 
theory of cell autonomy. 

I . The influence of mechanical pressure may be traced both in the incep- 
tion and in the course of tumor growth. In the former case it is somewhat 
difficult to visualize the action of pressure. Thiersch thought that relief of 
mechanical pressure followed the decay of the stroma about epithelial cell 


groups and permitted the expansion of the cell mass, but this element is 
inadequate to explain the inception of any tumor. The rapid lateral growth 
of regenerative epithelium in skin grafts may be referred in part to the ab- 
senceof lateral pressure. Loeb has made an effort to analyze this element 
experimentally. Given a certain grade of epithelial overgrowth, as in the 
tense ducts of chronic mastitis, trauma would reheve tension and facilitate 
expansion of the cell mass. Acting upon blood-vessels pressure is a promi- 
nent factor in controlling the progress of tumors. Fibrous encapsulation 
causes some tumors to regress, and rehef of pressure by incision of capsule 
releases active growth. These events are frequently seen in both early and 
late stages of tumor growth. 

2. Distribution of Nutriment. — Many chnical and experimental data have 
established the importance of blood- supply both for the physiological pro- 
liferation of cells and the inception and progress of tumor growth. The 
normal development of organs is regulated partly through the plan of the 
arterial system, and partial agenesia follows arterial defects. Cohnheim 
believed that cell rests begin to grow chiefly because of increased blood- 
supply. The minute study of many early tumors indicates that increased 
vascularity precedes the hypertrophy and multiphcation of cells. Yet rela- 
tive anemia is perhaps equally common. 

While the blood-supply is often increased the abnormal capacity of tumor- 
cells to absorb nutriment is a more obscure and significant feature of the 
disturbed nutrition of neoplastic cells. Its explanation would throw much 
Hght on the nature of the tumor process. The increased demands for func- 
tion determine increased vascularity in certain tumors (thyroid cancer). 
In the case of chorioma an increased supply of lutein secreJ;ion, according 
to Pick and others, is associated with abnormal growth of the chorion over 
whoso normal growth this secretion presides. It is possible that other hor- 
mones may by unusual activity favor the growth of tissues with which they 
are concerned, especially in the field of multiple tumors of paired or func- 
tionally related organs. Thus there is extreme fibrosis of the ovaries in 
many cases of mammary cancer. The theory that a specific nutritive sub- 
stance, not itself the nutrient molecule, presides over the growth of tumors 
has been apphed by Ehrlich to explain the results of his zig-zag transplanta- 
tion of mouse tumors, from mouse to rat and back. While this theory 
seems out of accord with certain observations concerning these transplant- 
able tumors and seems to have little bearing upon the problem of tumor 
growth in man yet it presents a new point of view for the study of the origin 
of tumors in general. 

3. The Influence of Specialized Functions. — The energies of cells are nor- 
mally divided between proliferation and specialized function, between 
work and growth. Normal function diverting the energies of the cells must 
be a constant restraint or regulator of growth. In most organs certain 
groups of cells are set apart for growth and from these are derived the more 
specialized functionating cells. The germ center cells of lymph-nodes, the 
cells at the bases of intestinal vilh, and the basal cells of the epidermis, are 
the ones that respond to demands for growth and from them tumors arise. 
Overresponse to demands for growth and loss of functional requirements 
both seem to precede the development of cancer, as in the adenomas of cir- 
rhotic fivers, and in the atrophying breast. Adami has presented the general 
importance of this principle designating the tumor process as " the cumula- 
tive habit of growth, replacing the habit of work." 

Thiersch and Beneke have interpreted the tumor process as a degenera- 
tive overgrowth of cells. That there is overfunction in some degenerating 


cells, as in the paralytic hypersecretion of glands, has been pointed out by 
Oertel who urges that overgrowth may also characterize degenerative proc- 
esses. In protozoa the germinative and vegetative functions reside in 
separate nuclear structures, the macro- and micronucleus, and it seems 
possible that the overgrowth of tumor-cells may be accompanied by excess 
of germinative nuclear material. This conception of idioplastic variations 
in the cells has been pursued by Rulf. 

Active growth is usually associated with active function. In thyroid 
cancer of fish there is apparently an example of a certain type of maUgnant 
tumor developing as a result of overresponse to functional stirnulus. There 
are many examples of adenomatoid hyperplasia from excessive functional 
stimuli and it seems probable that this element may be concerned in the 
inception of certain tumors. Proliferating chroma'tophores and mucous 
goblet cells exhibit excessive' function. Thus both overfunction and loss of 
function seem to be related to tumor growth. 

4. Organization. — The limitless growth of tumors early suggested that 
the cancerous tissue had lost the control of the organism. Recamier in 1829 
definitely formulated this idea and the conception has steadily enlarged with 
the progress of biology. 

The fertilized ovum reproduces its kind because of very complex factors 
which are best interpreted as the influence of environment on hereditary 
characters. Yet after the last analysis there is always an unexplained resid- 
uum about regeneration which justifies the teleological concept. The forms 
of regeneration among the lower animals show many processes that seem to 
be purposeful. The regeneration of limbs at the breaking joints of crusta- 
ceans; the variable regeneration of one or two heads in flat worms according 
to line of incision; the regeneration of the lens in Triton indifferently from 
capsule, iris, choroid, or retina; the polarity of the hydroid Antennularia, 
which transforms twigs into roots according to its accidental needs; illustrate 
the extreme adaptability of regeneration to the uses of the organism. The 
mechanism by which these results are brought about, whether purposeful or 
not, is called organization. Some biologists, notably Morgan, claim that this 
influence can be defined only in terms of purpose, and cannot be whoUy 
resolved into any known physical forces. However that may be, it is a fact 
that a tumor can be defined only in terms of purpose. All the known features 
of a malignant neoplasm are exhibited by the normal chorion, even the de- 
structive invasion and cachexia, but the process is not a tumor because of 
its purpose which shows that it is under the control of the organization. 

Embryology and cellular biology have thrown much light on the nature of the forces 
concerned in organization. 

First, it has been shown that at all stages the organization has an equilibrium which is 
maintained by the altruism of the tissues. The dividing embryo in some animals up to 
the sixteen-cell stage yields cells any one of which if separated from the others is capable of 
producing an entire embryo (Wilson, Zoja). Beyond this stage and usually before, the 
cells become differentiated, different cell groups going to produce different specific tissues. 
According to Weigert and Roux the differentiation results from an unequal division among 
the cells of the properties of the ovum (the mosaic theory). According to others (Spencer, 
Hertwig), the differentiation is accompanied not by loss but by suppression of certain 
properties of the ovum. Owing to the unequal distribution or activity of qualities or 
potencies, some cells come to possess more of one type and others come to possess more 
of another, while the sum total of all the potencies is necessary to maintain the equilibrium 
of the organization. Hence the cells, tissues, and organs stand in a relation of antagonism 
or correlation which maintains an equilibrium, any change in one affecting more or less all 
the others. This interdependence has been called by Hansemann, the altruism of the cells. 
Removal of one kidney is followed by compensatory hypertrophy of the other, through the 
physical factors of increased urea in the blood, increased blood-supply and nervous stimulus, 
all acting harmoniously under the influence of the organization. There is altruistic hyper- 



trophy of bones with tumors of the hypophysis, and altruistic atrophy of the adrenals in 
anencephahc monsters. A definite chemical substance conveys this altruistic influence in 
corpus luteum secretion which controls the growth of the chorion. 

Many recent observations and experiments have emphasized the importance of the 
current of nutrition in determining the growth of normal and neoplastic tissues. Murphy 
finds that the presence of a small fragment of growing splenic tissue prevents the growth of 
tumor-tissue implanted in the chicken egg, a result probably signifying that the growing 
splenic tissue diverts and absorbs the available nutriment. This principle has received 
experimental confirmation in an interesting way by J. Loeb in observations on plants. 
Stockard shows how the rate of regeneration of the amputated arms of the starfish depends 
on the number amputated. There are many clinical observations indicating that the 
growth and even the inception of tumors may depend chiefly on the stream of nutrition to 
the part, and also upon the character of the nutrition. 

Second, the study of cell division reveals a mechanism by which the altruism of the 
cells and the control of the organization may be disturbed. As long as the cells multiply 

Fig. 10. — J. Loeb's experiment with leaf of Bryophyllum calicynum, illustrating growth 
restraints. The leaf, notched as above, fails to develop shoots while attached to stem, but 
does so when severed from stem. The influence is thought to depend on the flow of nutri- 
ment, which ceases to pass from leaf when it is severed. 

by normal mitosis and each stage of differentiation in one group is accompanied by equiva- 
lent antagonistic qualities in others, equilibrium is maintained. But if inflammation or 
chemical or mechanical factors derange the mitotic process single chromosomes may be 
destroyed and the resulting cells will not have the proper antagonists to balance it. By 
repetition of this process a new type of cell may arise which is more and more removed 
from the normal and fails to receive any restraining influence from the organization. 

The study of tumor-cells shows that irregular, asymmetric and multipolar mitoses 
and destruction of chromosomes are frequent in malignant tumors. Hansemann applies 
the term anaplasia to the condition of such cells, which signifies loss of normal differentia- 
tion, specific function, and organization. These physiological properties which charac- 
terize malignant tumor-cells are usually but not always associated with pronounced changes 
in morphology. Anaplastic cells are not embryonal cells but a new type which have lost 
their place in the organization. More or less anaplastic cells occur in inflammation, but 


there are many degrees of anaplasia, and its occurrence in inflammation accords with the 
fact that inflammatory hyperplasia may pass into neoplastic. 

The abnormal capacity for growth has long suggested that tumor-cells 
have some relation to sex cells. In some animals the entire series of sex cells 
from the fertihzed ovum up to the new egg cell has been traced as a distinct 
series apart from the somatic cells. In this series the mitotic nucleus exhibits 
only one-half the usual number of chromosomes, and these instead of assum- 
ing a V shape and radial arrangement are ring or loop shaped or composed of 
coarse granules. This gametogenous mitosis has also been observed in the 
growing edges of tumors (Farmer, Moore, Walker) and it may be produced 
by chemical irritants. Its occurrence does not signify that tumor-cells are 
equivalent to sex cells, and yet in connection with the invasive properties and 
striking altruistic relations of both classes it shows that tumor-ceUs and sex 
tells have some interesting points of resemblance. Spencer, Hertwig and 
others claim that the sex potencies of somatic cells are not lost but only 
suppressed. On this basis WiUiams has built an elaborate and ingenious 
argument to show that tumor growth signifies the reawakening of the re- 
productive functions in the somatic cells. According to this theory tumor 
growth is a form of agamogenesis comparable to the budding of plants. 

Many have supposed that tumor-cells are fertilized cells, through conju- 
gation with leukocytes (Klebs), by parthenogenesis (Waldeyer), by conju- 
gation of endothehum and fibroblasts (Recklinghausen), by nuclear conjuga- 
tion (Auerbach, Bashford), or by endogenous cell infection. 

None of these hypotheses has survived criticism and the theory of cell 
autonomy remains content in the position that tumor-cells emancipated from 
growth restraints are merely exhibiting their natural capacity for growth. 

The theory of cell autonomy appropriates all the observations regarding 
the mechanisms of the isolation of ceU groups. It also accepts the principle 
that adult cells which have not lagged behind in development and are not 
embryonal or superfluous may be isolated and freed from growth restraints. 

Ribbert lays great stress on the mechanical separation of epithelial cells 
by new connective tissue. In early carcinoma of stomach, epitheUoma of lip, 
and psoriasis linguse he finds strands of new connective tissue growing 
between the epithelium and snaring ofi cell groups. Or, round-cell infil- 
tration of inflammatory origin beneath the epithelium creates an abnormal 
environment, and nullifies the restraints to grow. Central enchondromas 
and many exostoses he refers to groups of cells isolated by irregular ossifica- 
tion as in rickets. Malignant tumors he thinks may develop from groups of 
cells snared off in the course of benign tumors or from implantations during 
removal, as of ovarian cystomas. So carcinoma develops from isolated cell 
groups in cirrhosis of liver, or contracted kidneys. 

Further it must be assumed that trauma isolates predisposed cell groups 
as in many bone-sarcomas. The evidence that malignant chorioma may 
develop from emboli of normal chorionic cells is practically conclusive (Mar- 
chand. Pick). 

In «-ray carcinoma there is chronic edema, round-cell infiltration, and 
fibrosis altering the relation of the epithehum long before the downward 
growth of these cells. In chronic mastitis multiple cancer nodules appear 
at minute points where the fibrosis is unusually dense and the disturbance 
of normal relations most marked. 

There is Httle doubt that the minute study of the conditions surrounding 
the early stages of cancer, as instituted by Ribbert, forms a strong support to 
the theory of cell autonomy, while its bearing on the early diagnosis and pre- 
vention of cancer is obvious. Yet these observations fail to show why the 


isolated cells so often regress and only rarely succeed in producing tumors, and 
it is clear that the mechanical isolation of cells is, like the embryonal charac- 
ter, only a predisposing factor in tumor genesis. To cover this defect various 
hypotheses have been maintained. 

I. The isolated cells have been altered and their growth tendencies 
increased by previous irritation. 2. There is a local predisposition to tumor 
growth. 3. There is a general predisposition to tumor growth. 

1. Billroth's dictum "Without previous chronic inflammation cancer 
does not exist," while subject to exceptions, is yet so generally true as to 
establish the great significance of chronic irritation as a factor in tumor gene- 
sis. Interpreted according to the principles of the theory of cell autonomy it 
is clear that successive generations of cells subjected to chronic irritation will 
suffer increasing loss of growth restraints with disturbance of their normal 
rate of growth, the size of their nuclei and cytoplasm, their blood-supply, and 
their relations to neighboring cells. All of these changes are observable in 
the preliminary stages of lingual and other cancers. In this condition it 
seems but a short step to downward growth of such cells when the cumulative 
process reaches a sufficient stage or when more abrupt mechanical isolation 
•of cells occurs. Whether these cells in their continued progress are merely 
exhibiting normal powers of growth or have acquired new powers constittiting 
them an entirely new biological series it is difficult to decide. Here it must 
be urged that the continued growth of cancer is an entirely different problem 
from its inception. Many new factors intervene to facilitate extension and 
metastases. It is by no means clear that malignant tumor-cells are more 
viable than normal cells. They are more readily autolyzed and in early cases 
tumor emboli enjoy less resistance to destructive forces than normal cells. 

For the inception of many tumor processes, chronic irritation acting upon 
normal cells seems to adequately explain the observed phenomena. Thus 
the cumulative effects of chronic irritation alter the nutrition and growth 
activities of cells so that when they become isolated they are not normal cells 
but are in a state of disturbed equilibrium with a tendency toward exag- 
gerated growth. Such conditions are commonly observed in the precancerous 

2. A local predisposition to tumor growth must be assumed to exist to 
account for the capricious development of many neoplasms. The factors 
which lead to the inception of tumors in one locality or individual often seem 
incapable of bringing this result under other circumstances. Many of the 
factors constituting this local predisposition form the basis of Cohnheim's 
theory, including (i) Isolation without misplacement of embryonal cell 
groups. (2) Persistence of cell groups and organs which normally regress 
after embryonal life. (3) Formation of superfluous cell groups. (4) Gross 
■abnormaUties in development of regions, organs, and systems. (5) Disturb- 
ance in the idioplastic development of cell groups without visible changes in 

In addition, Rindfleisch assumed that imperfect nervous control is one of 
the factors predisposing to exaggerated growth of tissues and tumor forma- 
tion. While the effectiveness of such a factor may readily be granted, there 
seem to be no observations in favor of its existence except the absence of 
intrinsic nerves in tumor-tissue. Likewise imperfect idioplastic development 
may be conceded as certainly favoring tumor growth but in the absence of 
definite criteria of such a condition this factor remains largely in the field 
of speculation and as a problem for future studies. A reasonable speculation 
also is that of Israel who supposes that tissue cells, like bacteria, in the course 
•of repeated or abnormal generations tend to vary and in some tissues they 


vary in the direction of overgrowth. Yet as normal cells do not vary in this 
way Israel is assuming an original local predisposition to overgrowth. Thus 
the well established elements in the doctrine of local tissue predisposition 
belong under the embryonal theory, while the others still remain seductive 

3. A constitutional predisposition toward tumor growth has long been a 
favorite field of discussion. There is a vast amount of hterature concerning 
the relation or antagonism between cancer and tuberculosis, syphilis, malaria, 
diabetes, insanity, arthritism, obesity, gall-stones, dermatoses, heart disease, 
mental disturbance, etc. The many interesting observations in this field 
seem to have wholly failed to throw any light on the etiology of malignant 
tumors, nor have they established any definite connection between cancer and 
the diseases mentioned. 

It has long been recognized that cancer is prone to develop in plethoric 
individuals. Williams sums up an elaborate argument with the conclusion 
that cancer occurs chiefly in the well nourished, the well-to-do, and the well- 
protected against infectious disease, that it is especially prevalent in peoples 
with whom the consumption of meat is high, and that it is a tax on civiliza- 
tion. Yet all of these factors may exhaust themselves in increased longevity, 
by which more individuals of weaker type reach the cancer age, and none of 
them seems to establish any definite predisposition to the disease. It is 
conceivable that overnutrition should facilitate overgrowth but it is not so 
clear why overnourished tissues should tend toward neoplastic instead of 
normal growth. 

The immense variation in the nervous and physical conformation of 
individuals suggests to Borst that there may be equal divergences for or 
against neoplastic growth of tissue cells. The quality of the body proteins, 
the general organization of the body, the relation of tissues and organs to the 
vascular and nervous systems may give rise to individual peculiarities which 
are of importance in the origin or failure of tumor growth, but there are few 
facts to support this speculation. At best it can only mean that the capacity 
for tumor growth is far more general than the occurrence of tumors. The 
existence of multiple tumors is more properly referred to definite multiple 
disturbances of development or to the existence of several recognizable ade- 
quate factors than to any general predisposition. 

The individual of cancerous type is described by Benecke as revealing a 
large heart and arteries, small lungs and pulmonary arteries, long and 
capacious intestine, and well developed osseous, muscular, and adipose tis- 
sues. Yet these are just the subjects most likely to reach the cancer zone, 
where the age factor becomes far more prominent than anything we know 
about predisposition. 

Imperfect balance between the sexual and other organs, the status of the 
thyroid gland^ and the chromaffine system, have been suggested' as forming 
the basis of a cancerous predisposition, but the observations in these fields are 
as yet quite inadequate to form the ground of an effective argument. Many 
of the above considerations may be adduced in favor of the hereditary 
theory of cancer. 

According to Williams the growth of malignant tumors is favored by 
an altered metabohsm of the body consequent on the decline of the sexual func- 
tions. He refers to the increased frequency of cancer after removal or 
destruction of the ovaries (Wells, Kratzenstein) , and in cases of congenital 
absence or defect of the sexual organs (Neugebauer). Sticker also found that 
50 per cent, of^a series of cattle and horses with malignant tumors were 


The thyroid gland exhibits in most adult cancer patients that tendency 
toward involution which belongs to advancing years, but that this involution 
is more marked in cancer patients has not been shown; nor has any relation 
been estabhshed between cancer and the various members of the chromaffine 
system. A small proportion of adrenal tumors is associated with malignant 
tumors in other organs (Williams, Adams, Guthrie). The usual integrity of 
the pancreas does not favor the theory of Beard that cancer depends on defect- 
ive pancreatic secretion. 

Thus although the facts of tumor growth seem to demand the assumption 
of a constitutional predisposition, the basis of this doctrine is as yet imper- 
fectly laid. Clinical observations, studies in histogenesis, and gross anatomy, 
and chemical researches fail to reveal why the same factors succeed in one but 
fail in another subject to produce tumors. The most suggestive hypotheses 
seem to be those concerned with overnutrition, the dechne of the sexual func- 
tions, and congenital disturbance in the idioplastic development of the cells. 
The constitutional element seems to be too subtle for ordinary methods of 
study, and to show itself most clearly in the relation of heredity to tumor 
growth, where, however, with certain exceptions, its practical importance 
diminishes almost to the vanishing point. 

Heredity. — Nothing about cancer is more generally accepted than its 
hereditary nature, and nothing is less satisfactorily proven. The evidence 
favoring this doctrine consists chiefly in records, some very notable, of cancer 
families, and in statistical studies of the incidence of the disease in the rela- 
tives of numerous cancer patients. Recently experimental evidence has 
been adduced in its favor. 

Cancer Families. — In 1837 Warren reported a family history in which the 
grandfather had cancer of the lip, while the son, his daughter, two sisters, 
and the daughter of one of them, all died of cancer of the breast. The most 
famous cancer family is that reported by Broca, 1866, of Madame Z, the 
details of which were furnished by a medical member of the family. Of 
twenty-six members, mother, children, and grandchildren, reaching the age 
of thirty, sixteen died of cancer of breast, liver, or uterus. Napoleon I, his 
father, one brother and two sisters, are said to have died of cancer of the 
stomach. Sibley observed cancer of the left breast in a mother and her five 
daughters. Korteweg saw cancer of the breast, and Paget cancer of the 
uterus in mother, daughter and granddaughter. 

Retinal glioma occurred in 10 of 16 children of healthy parents in a family 
encountered by Newton, and in eight children of a family seen by Wilson. 
These, however, are instances of congenital, not hereditary origin. Williams 
has collected many other references to cancer families. 

Levin analyzes the conditions under which the clinical study of heredity 
in cancer may be conducted and points out that the occurrence of cancer in a 
family member may not necessarily mean an hereditary influence. An 
extended investigation of the families of five cancer patients failed to show 
any unusual incidence of the disease. In fact he seems to have encountered 
unusually resistant families in which cancer was of occasional but rare 
occurrence. Warthin on the other hand records several very striking 
histories of families in which cancer was unusually frequent. Thus in one 
family, 17 of 48 descendants of a cancerous grandfather developed cancer, 
chiefly of uterus or stomach. In another, eight descendants of a cancerous 
great-grandfather all died of the disease. Warthin adds that these families 
became extinct ; but clearly not from cancer. These observations emphasize 
the occasional occurrence of cancer families, and suggest that members of 
such families may properly take definite precautions against the disease. 


Statistical studies began with Paget who traced cancer or other tumors in 
the relatives of 23.6 per cent, of 254 cancer patients, but in only 18.3 per cent, 
of 147 non-cancerous tumor cases. Baker found cancer in the relatives of 
24.2 per cent, of 322 cancer cases. Velpeau thought that at least one-third of 
his cancer patients had a family history of cancer. Leichtenstein collected 
histories of 1137 cancer cases in which heredity appeared in 17 per cent. 

The proportion of women with cancer of breast in whose relatives there 
was a history of cancer has been estimated by Buthn at 37 per cent., by Nunn 
at 29.3 per cent., by Williams at 24.2 per cent., and by Leaf at 23 per cent. 
Yet Campiche and Lazarus-Barlow place the proportion in 100 recent records 
of the Middlesex Hospital at 15.6 per cent., Wilhams found a history of he- 
redity in cancer of uterus in 19.7 per cent., of breast 24.2 per cent., of other 
female cases 23.9 per cent., and in men suffering from cancer (209 cases) 11 
per cent. 

On the other hand Lebert placed the percentage of inheritance at 8.5 per 
cent.; Sibley in 305 cases at 8.5; Winniwarter in Billroth's clinic at 8.8 per 
cent. Billroth found one parent affected in 5.8 per cent, of 170 cases. TiU- 
mann quotes reports showing 8.5 per cent, and 10 per cent, of hereditary 
cases of cancer of stomach. Ziel collected 200 cases with 1 1 per cent, heredi- 
tary. H. Cripps excluded distant relatives, as cousins and aunts, and in 169 
cases only 6.4 per cent, had a cancerous parent. He found that in England 
and Wales (1861-70) i in 29.1 deaths occurring after 20 years of age was 
from cancer. Comparing this ratio with that of cancerous parents of cancer 
cases he concluded that cancer in one parent does not increase the liability to 
the disease. Recently Hillier and Tritsch made an elaborate study of family 
history in 3000 cancer cases in the Middlesex hospital. Of these 13. i per 
cent, had a history of cancer in relatives, while of 417 non-cancerous cases 
14 per cent, gave cancerous antecedents. Pearson who interpreted these 
data hesitated to accept the necessary conclusion that a family history of 
cancer confers a very sUght immunity to the disease, yet unless the collection 
of the data was very much biased it is clear that statistics do not support the 
doctrine of the general hereditary nature of cancer. 

Bashford points out that in England of those living at 35 years of age one 
man in eleven, and one woman in eight, die of cancer. This ratio signifies 
that a history of cancer will occur in one out of every two famihes of the 
general population. In the families of 669 cancer patients he obtained a 
history of cancer in 31 1 or 50 per cent. In Stuttgart, Weinberg examining the 
cancer records from 1872 to 1902 found no evidence that heredity plays a 
dominant part in the etiology of cancer. His data show that the deaths from 
cancer in each 100 relatives of cancer patients were, of the patients' parents, 
6.6, brothers and sisters, 3.9; of their husbands and wives' parents, 5.9; 
brothers and sisters, 3.1. 

There are numerous sources of error in the conclusions drawn from 
statistical studies. It is quite unhkely that any hospital population can give 
reUable information regarding the existence of cancer in their parents, much 
less in their other relatives. There are numerous errors both of omission and 
commission in the diagnosis of cancer. The nature of any hereditary element 
in cancer is wholly obscure so that it is impossible to decide whether cancer 
in cousins and aunts, or only in parents should be considered, whether the pre- 
disposition may skip generations, whether the influence affects the same 
organ or all the organs, or whether the influence of a similar environment, 
cKmate, food, and habits, may not submerge any form of heredity. Con- 
sideration of the age factor completely altered the conclusions suggested by 
Guillot's statistics. In 352 patients he found a cancer history in 10 per cent. 


of the non-cancerous and in 17.4 per cent, of the cancerous. But the average 
age of the Hving relatives of the non-cancerous was 37 years, that of the 
others 52 years. So that the expectation that cancer would later develop in 
the relatives of the non-cancerous before their average age reached 52 
changed the true percentage of cancerous relatives of the non-cancerous to 
16, or 1.4 per cent, less than for the cancerous group. In view of all these 
considerations it must be confessed that the statistics of cancer heredity in the 
present state of our knowledge are inconclusive or even worthless. 

The evidence presented by the many notable cancer families seems to have 
a different value. It seems impossible to attribute to coincidence, or envir- 
onment, or erroneous observation the remarkable prevalence of cancer in the 
famihes mentioned by Warren, Broca, Manichon, Le Tulle, and others. 
These cases seem to show that in rare instances a pronounced hereditary pre- 
disposition to the disease exists. It is, however, important to emphasize that 
the history of these famihes indicates that when an hereditary predisposition 
exists it displays itself in unmistakable form and it in no way justifies the 
assumption that hereditary influences prevail for cancer in general, on which 
point the evidence is wholly negative. These family histories show how very 
uncommon an effective hereditary influence really is. From the clinical and 
statistical evidence therefore one must, I think, subscribe to the conclusions 
of Le Doux-Le Bard, (i) Nothing authorizes us to affirm that cancer is 
hereditary. (2) In the interests of the public this doctrine ought to be 

The influence of heredity upon tumor incidence has been submitted to 
experimental test. Bashford bred mice of cancerous stock and ones that 
had actually suffered from the disease, attempting by inbreeding to inten- 
sify the hereditary influence. His results followed the laws of age and sex 

Tyzzer has conducted breeding experiments with the offspring of a female 
mouse, suffering from a cystadenoma of the lung. Of 100 offspring of this 
one cancerous parent 65 survived six months and 3 more than two and a 
half years. Of these, 20 (32 per cent.) developed one or more tumors, 17 
being pulmonary tumors. Among 68 of these mice born of cancer-free par- 
ents 9 (13 per cent.) developed tumors, while among 29 with one cancerous 
parent 1 1 (39 per cent.) developed tumors. The increasing frequency of tu- 
mors in the old mice (-2 years) of this family indicated that cancer was almost 
inevitable for animals which survived infectious diseases. While no satis- 
factory controls were secured these interesting results strongly support the 
hereditary theory, at least for this family of mice. 

Maud Slye has conducted elaborate experimental studies on the influence 
of heredity on the incidence of mouse tumors and has proven that inbreeding 
of tumor-bearing animals greatly increases the incidence of tumors. This 
work has continued over several years and includes over 10,000 autopsies 
revealing 722 tumors. It should be noted, however, that this hst of tumors 
includes and attributes equal significance to processes which are very distantly 
if at all related, such as lymphosarcoma, pecuhar tumors of the lung, and 
cancer of the stomach or breast. Until the experimental studies have been 
confined to identical or closely related neoplastic diseases, their results for 
mice must be regarded as of uncertain significance while the application of 
the results to man seems premature.' 

A general survey of the work in this field suggests several considerations 

which seem to deserve attention before definite conclusions can be reached. 

There are many indications that the conception of all tumors as a single 

' Similar results were later obtained with hepatic cancer alone. 


disease and the merging of all varieties in the search for signs of heredity 
are unwarranted, and that the studies should be confined to single closely- 
related tjrpes of malignant tumors in which common exciting and similar 
predisposing factors exist. Cancer of the breast in three generations of 
women is much more suggestive of heredity than a series in which mammary 
cancer is followed by hypernephroma, and this by teratoma ovarii. 

The Mendelian characters noted in the heredity of some pathological 
conditions have not been traced with tumors. It has often been urged that 
the cancerous predisposition may result from factors acting in each case in 
utero, i.e., that the predisposition is congenital without being hereditary, 
but this possibility has received little attention from the advocates of the 
heredity doctrine. 

It would be interesting to know if cancer in markedly hereditary sub- 
jects pursues an unusual course. In the absence of data on these and other 
matters and owing to our lack of knowledge of the nature of any transmissible 
influence, the problem of heredity in cancer seems for the present unsolvable. 


While one general and essential principle, best portrayed in the theory 
of cell autonomy, seems to underlie the inception of all tumors yet it is be- 
coming more and more apparent that the etiological factors are variously 
combined for nearly all neoplasms. In this sense it may almost be said that 
the etiology of each form of tumor is specific and that a practical knowledge 
of tumor etiology requires a minute analysis of all the factors concerned in 
each case. 

Thus the conditions surrounding the inception of cancer of the breast are in 
many respects quite different from those concerned with cancer of the tongue. 
This distinctive quality is even more obvious in the clinical course and 
termination of different malignant tumors, so that there is theoretical basis 
for and practical value in considering each tumor as a distinct clinical entity. 
Well-known examples of specific etiology are seen in various tumors connected 
with occupations, as :v-ray cancer, multiple epithelioma of the skin of 
workers in paraffin, and anilin dyes, scrotal cancer of chimney-sweeps, and 
the rare cancer of the lung in metal workers, weavers and cigar makers. 

The cumulative effects of the habits of the individual are seen in smokers' 
cancer of the lip, pharynx, and larynx, Kangri and Betel-nut cancer of African 
natives, and cancer of the skin in arsenic eaters. 

Particular forms of chronic irritation and inflammation precede many 
well-known clinical forms of cancer, as epithelioma of tongue from jagged 
teeth, cancer of skin from chronic eczema, indolent ulcers, burns, or along 
chronic sinuses from necrosing bone, and with Paget's disease of the nipple. 

Specific chronic granulomas occasionally lead to cancer, as syphilitic 
psoriasis linguae, and lupus of the skin. Sarcoma also develops after lupus 
and syphilis of bones. Chronic inflammation of mucous membranes pre- 
cedes cancer as in uterus, stomach, urinary passages, nares, and with chole- 
lithiasis. Chronic inflammation of glandular organs frequently forms the 
essential basis of cancer, as in the cirrhotic liver, contracted kidney, and hy- 
pertrophic prostate. Chronic hyperplastic and benign neoplastic processes 
rather rarely become transformed into mahgnant tumors, as warts, polj^ps, 
myoma uteri, and simple epithelial cysts of many organs. Abnormal in- 
volution designates the conditions leading to many cancers of the breast, 
appendix, prostate, and thyroid. Trauma seems to be the sole tangible 
factor in originating many tumors of bone, glia tissue, testicle, and many 
other organs. 

The wide differences in the nature of the embryogenic disturbances which 
lead to isolation of embryonal cell groups and tissues set apart a very large 
class of tumors whose etiological factor is distinctive for each member as 
well as for the entire class. Thus branchiogenic cancer, Wilms' renal em- 
bryoma, and the teratoid testicular tumors which are probably derived from 
sex cells, deserve to be considered as enjovjilg quite a different etiology. 

If, as seems probable, Hodgkin's.g the new vtSfiomes transformed into 
a nearly autonomous lymphosarcon '■ttle resemblance be- 

tween its etiology and that of bi\ls about fractures reveJPi the testis of a 


child. If abnormalities of the lutein secretion enable emboli of normal 
chorionic cells to develop maUgnant tumors there is little wisdom in discuss- 
ing its etiology with that of smokers' cancer of the larynx. The two dis- 
eases are as different in etiology as gout and erysipelas, both of which pro- 
duce exudative inflammations of the skin. 

Thus the special etiology of tumors presents extremely numerous and 
complex problems which have immediate practical bearing on diagnosis, 
treatment, and prevention, and which are probably much nearer solution 
than is that of the nature of the neoplastic process. It seems iniportant to 
emphasize the separate nature of these departments of the subject of eti- 
ology and to point out that in many respects our knowledge of the etiology 
of tumors is very clear and precise. Yet the action of all these special etio- 
logical factors, as irritation, trauma, and inflammation is only secondary 
and indirect, and without combination with other predisposing conditions 
they are incapable of inducing tumor growth. In a great many cases the 
above special factors exist but no tumor develops. 

Therefore the special etiology of tumors can deal only with the conditions 
under which tumors arise, but not with the intrinsic properties of cells which 
find their expression in the tumor process. 

Trauma. — ^Mechanical trauma is an important factor in the causation 
of tumors. From the scope of trauma in this relation should be excluded the 
various forms of chronic irritation, inflammation, surgical wounds, cicatriza- 
tion- and chemical escharotics. By trauma is here understood a single or 
repeated more or less contusing, crushing, or lacerating mechanical injury. 
This much discussed subject would be greatly simplified if the statistical 
tendency were replaced by clearer ideas of the results of trauma in different 
tissues and of the mechanism by which such lesions lead to tumor growth. 
Studies in this field are as meager as the statistical contributions are 

The important effects of trauma here are: (i) Solution of continuity, 
minute and gross; (2) separation of cell groups and tissue masses, as of skin, 
glands, bone; (3) necrosis of tissue; (4) confined hemorrhage requiring ab- 
sorption or encapsulation; (5) accelerated regenerative processes with hy- 
peremia, and newgrowth of specific cells, blood-vessels, and supporting 
tissue; (6) cicatrization. 

All of these conditions are well-proven elements entering into the causation 
of tumors, and the failure of attempts to produce tumors by experimental 
trauma in given cases does not reduce their importance when associated 
with other necessary predisposing factors. When one or several of them 
meet with the relatively rare predisposing conditions, we have a somewhat 
satisfactory explanation of the facts observed regarding the relation of 
trauma to tumors. 

The predisposing factors take many forms: (i) There may be a benign 
or a minute malignant tumor in the tissue before the injury. Many patients 
with cancer of the breast attribute their disease to some form of injury. Prob- 
ably very few of these tumors are the direct sequel of the trauma, but a slowly 
growing cancerous nodule in chronic mastitis may be accelerated by a blow 
where the injury alone seems to be the immediate cause of the tumor. In 
any organ a preexisting lesion renders the effects of ordinary injury more 
severe and more noticeable. 

2. The precancerous rr^-^"' ' -nay be precipitated into a malignant 
process by injury. E- ids of a psoriatic tongue by the teeth, 

injuries of the br*- 'c mastitis, and incomplete surgical 

removal of ind ' ^s, fistulous tracts, and benign 


tumors. Leukoplakia of the tongue is said never to develop cancer until it 
becomes complicated by cracks and fissures. 

3. Misplaced and undeveloped organs are predisposed not only to tumor 
growth but also to trauma. The results of this unusual combination are 
seen in cancers of undescended testes and supernumerary breasts. 

4. Aberrant quiescent cell groups may be included in the damaged tissue. 
The best known example is the malignant melanoma arising from an injured 
mole. Definite injury frequently precedes the appearance of the various 
forms of teratoma testis, and since trauma is an effective method of producing 
artificial parthenogenesis, there is good reason to believe that the relation 
of the injury to the tumor is in this case direct, for these tumors develop 
from immature aberrant sex cells. Throughout the entire series of embryonal 
tumors there is a sound basis for ascribing more than ordinary significance 
to a history of severe or mild and repeated injury. Ribbert thinks that many 
adrenal rests are incited to growth by trauma. 

5. Normal cells under the conditions established by trauma develop 
benign or malignant tumors. There seems to be no sufficient reason for 
denying the relation between trauma and many chondromas, osteomas, 
lipomas, fibromas, and fibromyxomas, with which the history of a blow is 
rather common, and there is little ground for assuming that the trauma acts 
on any but normal cells. Also with the two most striking forms of traumatic 
tumors, gliosarcoma and osteosarcoma, where a definite history of injury 
is frequent, there is no evidence that any but normal cells are involved. 

The character of the injury followed by tumor growth varies widely. 
Traumatic cancer usually appears after repeated mild injuries often amount- 
ing to chronic irritation, while sarcoma commonly develops after a single 
blow. Yet many cancers have appeared after single injuries (Ziegler, 
Lowenthal, Jordan, Second). Only the severer injuries are capable of reach- 
ing the deeper tissues from which sarcoma develops. Cancer has been clearly 
traced to lacerations by rough instruments, rusty nails and pins, thorn pricks, 
insect bites, surgical wounds, and to blows without visible destruction of 

Fractures, lacerations of deep tissues, contusions and concussion, all of 
which are apt to cause hemorrhage, represent the usual type of injury pre- 
ceding sarcoma and benign mesoblastic tumors. 

The mechanism by which injury induces tumor growth is generally 
obscure. The traumatic dermoid clearly originates from implanted epider- 
mis and fair success has attended its experimental production. Yet it has 
been abundantly proven that simple misplacement of other normal tissue 
cells is quite inadequate to explain tumor growth. Hyperemia may be the 
remaining factor necessary to excite the proliferation of aberrant rests, and 
precancerous foci. Blood extravasation requiring absorption or organization 
creates conditions especially favorable to mesoblastic overgrowth. Certain 
tissues are predisposed to excessive overgrowth in the presence of hemorrhage. 
The periosteum, the tissues of the scrotum and cord, the omentum, and the 
pia-mater frequently exhibit very active overgrowth of tissue following 
hemorrhage, yielding bulky products, which have the gross and many of the 
microscopical appearances of tumors. Such pseudosarcomas are especially 
frequent in the scrotum. Cornil and Berard have noted the very active 
proliferation about traumatic hematomas, and locate the early sarcomatous 
process in the cells about new blood-vessels. Some early sarcomas of the 
brain present appearances implicating the new vessels in the sarcomatous 

The common superfluous callus about fractures reveals the great pro- 


liferative capacity of periosteum, and Benecke believes that the regenerative 
process in periosteum may run directly into sarcoma. In traumatic myositis 
the new tissue is often suggestive of sarcoma. Yet the full explanation of 
traumatic sarcoma seems to require the assumption of special local or general 
predisposing factors which have never been traced in microscopical structure. 
The chemical composition, blood-supply, and lymph supply of brain tissue 
are peculiar but give no clue to the frequency of traumatic ghosarcoma. The 
presence of necrotic tissue should receive attention as a possible specific 
factor in the development of traumatic tumors. It is possible that products 
of autolysis or decomposition of protein tissue may exert peculiar influence 
on regenerative processes, as in the brain. Kottnitz cites many cases of 
lipoma in which a traumatic factor appears acting directly or through the 
nervous system. 

Ribbert admits that we have not the sUghtest knowledge why trauma, 
which affects many cell types in large areas, excites tumors of one cell type 
in one focus. He thinks that the results of experimental separation of cell 
groups have often approached the condition of benign tumors some of which 
may regress. The most extensive proliferation is obtained by partial separa- 
tion of tissues, as along three sides of a square, of the rabbit's ear. EpitheHum 
completely separated dies, but regenerating connective tissue is much more 

The frequency of the traumatic origin of tumors varies with different 
statisticians, according to their conception of trauma and their critical 
standards. Composite statistics vary from 2.5 per cent, by Kempf to 
44.7 per cent, by Lowenthal. Of 6780 tumors of all varieties reported in the 
literature with reference to trauma 494 or 7 per cent, were regarded as of 
traumatic origin. Of 2641 cancers, 107, 5 per cent., and of 938 sarcomas 186, 
or about 19 per cent., were classed as traumatic. In a large proportion of 
these cases the relation of the trauma was extremely uncertain. The figures 
represent the present attitude of most cHnical writers which, as Lubarsch 
and Schimmelbusch have said, is often extremely uncritical. Only the 
statistics relating to single tumors are of practical value. Of 55 melanomas 
Werner-Rowe found 19 with a history of injury of a mole. 

Of ghoma of the brain Adler collected 1086 cases of which 8.8 per cent, 
were preceded by rather definite traumatic history. In this class of cases 
the relation between the trauma is often clear, the symptoms of trauma 
merging into those of tumor and occasionally a scar has been found in bone 
or dura directly connected with the tumor. 

There are important medicolegal aspects of the relation of trauma to 
tumors. To establish the relation of trauma several classes of evidence are 
necessary, as is recognized especially by the French statutes. 

I. The authenticity and sufficient importance of the trauma. (2) 
Previous integrity of the wounded part. (3) A reasonable time relation, 
three weeks to three years or more in certain cases. (4) Continuity of patho- 
logical changes or symptoms in the wounded part and the appearance of 
the tumor. (5) Microscopical proof of the existence of a tumor. 

It is necessary to distinguish also between, complete and partial liability. 
When a healthy adult receives a blow on the head, remains unconscious for 
days, never recovers full mentality, shortly develops signs of brain tumor, 
and dies in a few months of gliosarcoma of suitable size and location, the 
liability is generally accepted as complete. When a patient with symptoms 
of abdominal disease receives a blow in the epigastrium and rapidly succumbs 
to carcinoma of an abdominal organ it is reasonable to assume that the 
trauma hastened the progress of an existing tumor. As Segond states. 


trauma reveals, aggravates, or serves as a pretext for the origin of tumors; 
so that in each case very careful consideration of all circumstances are 
required to establish the relation of the injury. 

In the German courts definite liability has been assumed for cases in 
which the connection between trauma and tumor is very uncertain and in- 
capable of proof, as with carcinoma of colon or stomach appearing several 
years after a blow on the abdomen. 


The parasitic theory is the oldest hypothesis of the origin of cancer. 
It appealed to the ancients, was tacitly accepted throughout the Middle 
Ages, was definitely argued by modern observers, and reached the height 
of its popularity as a scientific theory about 1895, but during the last fifteen 
years it has rapidly lost ground, and today few competent observers consider 
it as a possible explanation of the unknown element in blastomatosis. 

The data concerning the parasitic theory fall into four main classes: 
(i) Studies of the incidence of cancer; (2) the search for specific parasites 
in tumors; (3) the experimental production of tumors; (4) theoretical 
consideration of the nature of cancer. 

The Incidence of Cancer. — In 1809 Arnaudet reported that in certain 
rural districts in Normandy during a period ,of eight years, the cancer mor- 
tality reached 14.88 per cent. He noted the occurrence of cancer in several 
members of the same household, and he concluded that the neighborhood 
of a cancer patient was contaminated through the water. 

Many similar reports of cancer districts and cancer houses soon appeared 
from Guelhot, Webb, Fiessinger, D'Arcy Power, Bosc, and others. Behla 
found the cancer mortality of the low-lying portions of Luckau to be four 
times that of the central portion. He noted the frequency of cancer in the 
dogs of the town, the large quantities of home-grown vegetables consumed 
by the cancer famiUes, and concluded that cancer is an infectious disease 
carried to its victims through contamination of the soil. 

In 1892 Haviland from an elaborate analysis of the English census 
concluded that cancer is a disease of low-lying districts and of seasonably 
flooded riparial lands where subsoil is constantly moist, while in high dry 
localities it is rare. 

Very minute studies of cancer locaHties have been made in England and 
other countries by numerous observers all tending to show a relation between 
cancer and all manner of local pecuHarities; as hving in "old and cancer 
houses" (Park), " infected streets " (Mason), houses with defective plumbing, 
leaky sewers (Nason), proximity to trees, especially large ones (Lloyd- 
Jones), collections of decaying vegetable matter, stagnant water (Poppel- 
man), abundance of certain insects (Bosc), etc. The uncritical character 
of these observations hardly needs comment. Many of the fallacies involved 
in them have been pointed out by Sticker and by Prinzing. 

At the height of their popularity, Symons showed that there were no 
cancer houses in Bath wher« cancer mortahty was the highest in England. 
Out of this fanciful discussion Williams draws the sane conclusion that cancer 
is a disease of civihzation, choosing its victims from the well-to-do, the well- 
nourished and the well-protected against infectious diseases, and that it 
flourishes in just the opposite conditions from those which favor the spread 
of tuberculosis. At one time there was a tendency to report small epidemics 
of cancer among men (Webb, Fiessinger), and to point out its epidemic char- 
acter. Later, epidemics of cancer among cattle were described (Loeb), 
or among rats and mice used for breeding purposes (Gaylord, Borrel, Loeb), 



for which old age and inbreeding are sufficient explanations. Bashford, in a 
very wide experience with captive rats and mice found no evidence of any 
"cage infection." 

Epidemics of thyroid cancer among artificially bred trout have menaced 
the fish industry in several countries. Pick and Plehne concluded that it 
arises from overcrowding, overfeeding, and inbreeding. Gaylord demon- 
strated the true cancerous nature of a certain proportion of these cases, but 
Gudernatsch pointed out the remarkably wide distribution of the thyroid 
lobules in many species of fish and rendered doubtful the neoplastic character 
of the great majority of the "tumors." Marine and Lenhardt have shown 
that many of the hyperplastic thyroids regress under proper hygiene. 

Statistics elaborately presented by WiUiams and by Hoffman show a 
remarkable increase in the recorded deaths from cancer during the last 
fifty years. In England in 1840 the cancer deaths were one in 129 of the 
total; in 1905 one in 17. Very similar indications are found in the records 
of most civilized countries (Hoffman, Lit.). Williams calculates that in 
England, 1900, i in 15 of all men, and i in 9 of all women, living at 35 years 
of age are destined to die of cancer. The increase seems to be affecting 
men more than women, although in England cancer is now a more fatal 
disease for women than is tuberculosis. The increase of cancer is con- 
comitant with a decrease of tuberculosis. 

Of the apparent increase in cancer deaths a considerable proportion must 
be referred to improved diagnosis. Riechelman shows for the Berlin hospitals 
that there was still, 1902, room for a 20 per cent, increase in the recorded 
deaths from cancer through improved diagnosis alone. The remaining por- 
tion must be largely referred to the longer tenure of life through immunity to 
infectioas diseases and improved hygiene, which permit a larger proportion 
of persons to reach the cancer age. In Prussia, vital statistics show that the 
average person hves twenty-five years longer than in i860. Willcox in a 
critical analysis of this subject shows that the sources of the recorded increase 
are quite complex, including improved diagnosis, changes in age composition 
of population, eUmination of old age as a cause of death, improved regis- 
tration, and widening of the definition of cancer. 

He shows that the increase is chiefly of inaccessible cancers, among the 
lower classes of society, in males rather than in females, in negroes and the 
foreign born, and in the country more than in the cities. 

The impression that cancer is actually increasing to a slight extent which 
cannot wholly be explained by the above factors and is more often appearing 
at earher ages, is probably correct, but such a fact cannot stand as an argu- 
ment for the parasitic theory. 

The contagiousness of cancer was at one time supposed to be proven by 
clinical observation. Lusitanus (1557) claimed to have observed infection 
of three children by a cancerous mother. Tulpius reported the direct 
transfer of the disease between human beings. In the 17th and i8th cen- 
turies cancer was regarded as quite as infectious as phthisis with which it 
was often confounded. Lebert and Friedreich believed they had observed 
cancer infection of the fetuses of cancerous mothers. 

The numerous observations on contact infection were supposed to point 
to a parasitic factor but relatively few of these cases seem to have been genuine 
and in these it must be held that the infecting agent is the cancer-cell. Of 
the many cases of cancer a deux reported by Budd and GueUiot none appears 
to be properly attested, while Demarquay found that in only one of 134 
cases of cancer of the penis did the wife have cancer of the uterus, and Bossi 
reported that of 180 husbands of women with uterine cancer, none contracted 


the disease. The attempts of Alibert, Wickham and Senn to inoculate them- 
selves with cancer, failed. Finally, it is the universal experience of surgeons 
that the disease is not contracted by the treatment and care of cancer patients. 
The conclusion may thus be drawn, that nothing in the incidence of 
cancer points to its infectious nature. 

Microorganisms in Cancer. — With the advent of microbiology each new 
class of microorganisms in turn was isolated from cancer tissue, many were 
grown in pure culture, and some were claimed to reproduce the disease upon 

(a) Bacteria. — Rappin, Schill, Francke, Lampiasi, and others made 
unwarranted claims for the data regarding the bacteria isolated by them and 
they were soon forgotten. In 1887, Scheurlen obtained pure cultures of a 
bacillus from mammary cancer and with it claimed to have produced tumors 
of the breast in dogs. Baumgarten showed this bacterium to be the potato 
bacillus. In 1890 Koubassoff offered a complete claim of evidence for a thick 
motile bacillus from gastric cancer. In 1910 Doyen announced his Micro- 
coccus neoformans as the cause and cure of cancer. It was soon shown that 
this microorganism had nothing to do with cancer, and the Societe de Chirur- 
gie (1905) reported it to be valueless in the ameUoration of the disease. 
These studies led to certain results of positive value . Shattuck and Ballance 
showed that non-ulcerated tumor-tissue could easily be kept sterile. Verneuil 
in necrotic foci and Zahn in some metastatic growths at autopsy found many 
bacteria. Richet found pyogenic bacteria common in cutaneous cancers. 
MaragKano often found staphylococci in the blood of patients with ulcerating 
tumors, even without fever, but never with non-ulcerating growths. 

{b) Coccidia, differing in many details, v/ere described b}- several observers. 
Darier's psorosperm of Paget's disease secured much support from Wick- 
ham and the histological evidence alone sufficed to impress many with the 
behef that this structure was the true cause of this and other cancers. It 
was repudiated by its discoverer in 1904. 

Thoma (1888) and Sjobring (1890) described much the same intracellular 
and intranuclear structure as a parasitic coccidium. 

Metchnikoff supported Sudakiewitsch's intracellular coccidium. 
Adamkiewicz asserted that all cancer-cells are parasitic coccidia ^Coc. 
sarcolytum) giving origin to a specific toxin (cancroin) and much this same 
view was held by L. Pfeiffer and J. Clarke. 

(c) Miscellaneous Protozoa. — Other observers could not regard all the 
cancer-cells as alien parasites, but certain of the tumor-cells they identified 
as parasitic amebse because of their bizarre forms and long pseudopodia 
which stretched between adjacent cells. These were the Rhopalocephalus 
carcinomatosus of Korotnefi, and Cancriameba macroglossia of Eisen. L. 
Pfeiffer described and depicted intracellular structures which resembled 
the microsporidia of muscle-tissue. 

Podwyssoski and Sawtchenko described as sporozoa a variety of free and 
encapsulated intracellular structures many of which resembled Sudakie- 
witsch's parasites. Rufifer and Walker improved the technical methods of 
demonstrating the cancer bodies and endeavored to distinguish between 
true and spurious parasites. Kahane thought he detected a minute proto- 
zoan in the circulating blood of cancer patients. In cancerous ascitic fluid 
Schaudinn observed a large ameboid cell which he named Leydenia gemmi- 
para. Schiiller traced the complete cycle of a minute intranuclear protozoon 
ill cancer-cells which differed from nearly all other cancer parasites. 

Bird's-eye Inclusion. — From the beginning of the search for the cancer 
parasite special interest always centered in a certain intracellular body called 
the "bird's-eye inclusion." 


This body appeared in the CN'cle of many of the parasites described by 
various authors. It was first cautiously suggested by Foa as the probable 
cancer parasite. In 1902 Feinberg made a final effort to establish specific 
features in this body, double contour, metachromatism and nucleus sur- 
rounded by a clear zone, but these features were promptly and authorita- 
tively discarded by Hertwig. 

[d) Mycetozoa. — ^L. Pfeiffer early pointed out the resemblance of many 
cancer bodies to forms of mycetozoa, especially Plasmodiophora brassicas. 
This organism was first described by Woronin as the cause of club root, a 
common disease of plants, and its complex cycle was worked out by Nawas- 
chin. Some of its forms closely resemble the bird's-eye inclusion of cancer, 
while others are extremeh' minute and difficult of detection in the infected 
cells. Behla found the disease common in the gardens of his cancer houses 
in Luckau. Elaborate attempts to demonstrate that this or a similar organ- 
ism is the cause of cancer have been made by Behla, Podwyssoski, Feinberg, 
Gaylord, and Robertson and Wade. 

Several observers claimed to have secured cultures of these protozoa 
and to have reproduced tumors by inoculation of the cultures in animals. 
Sjobring used a medium containing ascitic fluid, peptone, glucose and soap 
made from human fat, and claimed to have isolated his rhizopod from human 
tumors. Cultures produced tumor-like growths in mice. Yet the German 
pathologists denied that the growths were neoplasms and Israel asserted 
that the rhizopods were fat droplets. By the inoculation of material from 
club root Podwyssoski and Gaylord produced granulomatous swellings in 
animal tissues, the cells of which contained the englobed parasites, which 
closely resembled some of the intranuclear bodies seen in cancer. Yet v. 
Tubeuf and others have inoculated many animals with- club root and failed 
to produce any lesions resembling neoplasms. According to v. Tubeuf the 
histology of club root is not that of a neoplasm, the tumor resulting from 
distension and degeneration of cells surrounded by an area of inflammatory 
overgrowth. The picture of infected cells in club root distended with 
enormous numbers of very definite parasites is wholly different from anything 
seen in cancer. 

Smith has shown on an elaborate scale that Bacillus tumefacieus, isolated 
from crown gall, is capable of producing a variety of tissue overgrowths in 
plants which he regards as true tumors, both simple and teratomatous. Yet 
this observer fails to distinguish between chronic productive inflammation 
in plants and neoplasia, he does not consider that even in animals inflamma- 
tory hyperplasia may pass into neoplastic, while his infectious "embryomas" 
in plants cannot have any relation to the embryogenic disturbances known 
to give rise to teratomas in the animal body. 

Schiiller claims to have cultivated his protozoon and to have produced 
by it both carcinoma and sarcoma. Yet no one else has been able to find 
this organism in cancer and Schiiller's technic feU under suspicion when 
Volcker pointed out that the very characteristic large forms of this protozoon 
exactly resemble cork cells. 

(e) Spirochete. — In 1905 Borrel reported the occurrence of spirochete 
in two mouse tumors without attributing to them etiological significance. 
Wenyon has shown that mice are susceptible to blood infection by these 
organisms. Gaylord and Calkins, finding a single type of spirochete 
in ten mouse cancers and in 16 transplants therefrom suggested that the spiro- 
chete were the cause of the tumors. Yet Gaylord and Clowes succeeded 
in freeing their tumors from spirochete by treatment with KCN, and 
Tyzzer found many tumors in mice not infected with spirochete. In 


thirty-five human tumors and twenty-five in the dog I found spiral organisms 
only on ulcerated surfaces or in necrotic areas. Similar results were obtained 
by Mulzer and by Lowenthal. 

(J) Blastomycetes. — In 1890 Russell fully described certain small and 
large round bodies in cancer-cells which stained intensely with fuchsin and 
which he recognized as parasitic budding fungi or yeasts. Although these 
bodies had long before been noted by Fox, and Klein and Lubarsch soon 
showed that they were found in many normal and diseased tissues, the yeast 
theory of cancer was taken up systematically by San Felice and the ItaUan 
school. From 1895 up to the present time San Fehce has devoted much 
labor to the support of this theory. He experimented first with a culture 
obtained from fruits, Saccharomyces neoformans, and one from the lymph- 
nodes of an ox dying with carcinoma of the hver; Saccharomyces litogenes. 
Later he obtained cultures from human tumors. With these cultures he 
produced infection of many animals, chiefly dogs. 

With Saccharomyces neoformans he thought he produced a sarcoma in the 
breast of a bitch and in many other animals, but most of the lesions were 
clearly inflammatory. Saccharomyces litogenes produced very similar lesions. 

Repeated passages increased the virulence of the strains and their neo- 
plastic properties. Cultures from human tumors were non-pathogenic for 
animals. It was not possible to recover the organism from the tumors, a 
result which San Felice attributed to the death of the yeasts which then 
assumed the form of fuchsin bodies. In his latest studies San FeUce com- 
bines the soluble toxines of yeasts with the living cultures, finding that the 
toxines rather than the bodies of the organism cause proliferation of cells. 
Nevertheless the lesions which he pictures seem to have the features of infec- 
tious granulomas and not those of tumors. 

Many other observers claim to have isolated blastomyces from tumors 
and produced other tumors by inoculation, as Plimmer, Roncali, Corselli 
and Frisco, Curtis, Monsarrat, Leopold, Wlaeff, Klein, etc. Against their 
conclusions stand an equal number of very competent studies which show 
that blastomyces are rarely present in tumors and that they are incapable 
of producing neoplastic lesions. Foulerton and Richardson examined several 
hundred human tumors without securing a single yeast. Their rare presence 
in tumors is attributed by most bacteriologists to secondary infection, and 
the high proportion of successful cultures secured by many observers is 
clearly the result of air contamination. Meser found lycopodium seeds 
from the dressings deep within the tissue of carcinoma. The pathogenic 
action of a number of yeasts have been elaborately studied by many com- 
petent observers including Mafucci and Sirleo, Foulerton, Rabinowitsch, 
Petersen and Exner, and Nichols, none of whom found any indication that 
these organisms can produce tumors. 

The various blastomycetes known to be pathogenic for man produce 
characteristic granulomas but not tumors. It is true that in certain chronic 
blastomycotic lesions, especially of the skin, considerable hyperplasia of 
epithehum resembling cancer has occasionally been seen, but there is no 
indication that this hyperplasia exceeds that observed with syphilis and 
tuberculosis. Genuine cancer sometimes follows such lesions and yet no one 
supposes that the tubercle bacillus is the cancer parasite. The identity 
of yeasts, dead or ahve, with Russell's bodies has been subjected to careful 
criticism by many observers, especially by Sternberg and Nichols, who 
conclude that there is no adequate proof that blastomyces produce any con- 
siderable proportion of the fuchsin bodies. 

Otto Schmidt claims to have isolated from human carcinoma and sarcoma 


an organism which he calls Mucor racemosus and which he believes falls 
in the class of mycetozoa. With this organism he has produced ten sarcomas 
in rats and mice, some of which he has transplanted through several genera- 
tions, the tumor strain finally reaching loo per cent, of successful implan- 
tations. He is able to immunize rats against the organism and against 
its tumor-producing properties, and rats with tumors he is able to cure by 
inoculation with killed cultures of the mucor. Six out of 19 cancer cases at 
the Heidelberg Institute gave what he regards as anaphylactic reactions 
against the mucor. Baisch has succeeded in verifying some of Schmidt's 
results. On investigation it does not appear how often Schmidt secures 
cultures from human tumors nor with what precautions the successful cul- 
tures were obtained. The mucor cannot be identified in sections of the 
tumors. Some of the growing forms of the parasite Schuberg regards as fat 
droplets. From the experimental tumors the parasite is not recoverable. 
Ten tumors were produced, but hundreds of animals were used and it does 
not appear that the proportion of successful results greatly exceeds that of 
spontaneous tumors in rats subjected to other traumatic influences. The 
cure of the rat tumors is accomplished after very severe intoxication by the 
injected cultures and the death of many animals. Schmidt claims to have 
cured human cancer by his toxins but others have not succeeded. There 
seem to be no grounds for accepting the conclusion that human patients show 
anaphylaxis to the cultures of mucor. In fact Schmidt has not adduced 
satisfactory evidence that Mucor racemosus has anything to do with human 
cancer. What he seems to have accomplished is to add to the hst of toxic 
agents that are known to excite hyperplastic inflammation which, especially 
in rats, may exhibit metaplastic changes and possibly pass over into tumors. 
He has not apparently controlled his work by equally energetic efforts to 
produce similar changes by other agents. Yet his observations seem to fall 
in line with those of other observers who by various means succeeded in 
causing notable hyperplasia and metaplasia of animal tissues, as Fischer 
with Sudan III, Jores with Scharlach R in paraffin, Stohr and Stoeber 
with naphthalamin, paratoluidin, and amidoazotoluol, and Stoeber and 
Wacker, with indol and skatol. With the expressed juice or the nucleo- 
proteids of yeasts Galeotti and Pentimalh have also produced a considerable 
variety of tumor-like hyperplasia in dogs and rats, but they hesitate to assume 
that these processes are identical with progressive neoplasms. It is even more 
doubtful if the agents they used actually occur in spontaneous tumors. 

The Experimental Study of Tumors. — The inoculabihty of carcinoma 
was once supposed to demonstrate its parasitic nature. Long after Peyrilhe's 
failure several experimenters claimed to have succeeded in transferring cancer 
to lower animals by intravenous inoculations of fresh emulsions (Langen- 
beck, Follin, Velpeau). These results were later met with many more nega- 
tive reports by Virchow, Weigert and Billroth. The more recent claims of 
successful implantations of human cancer into lower animals by Boinet 
(1894) (epithelioma into dogs), Juergens, 1896 (melanoma into rabbits); 
Dagonet, 1904 (epithelioma into a rat); and Werner, 1907 (carcinoma into a 
dog) leave open the question how far human cancer may survive in these 
animals. Extensive melanosis of the organs as well as local tumors followed 
the inoculations of human melanoma by Goujon, Lang and Bosc, and Vedel, 
while Pfeiffer thought he was able to carry his melanoma into one of a second 
series of mice. Yet Fischer repeatedly failed to transfer melanoma to rats 
and Roux and Metchnikoff were unsuccessful with the chimpanzee. In 
none of the successful cases does it appear that the resulting process was 
anything more than a local or general reaction to the cells and pigment intro- 


duced. While it is possible that the reactions observed in some of these 
animals must be classed with the tumor-like processes produced by chemical 
■agents acting upon tissue cells there is no proof that the injected human cells 
multiphed during the considerable periods over which the animals were 

Firket, Boinet, Gaylord, Lewin, Bosc and Vedel, and others, claim to have 
produced cancerous nodules by inoculation into animals of various human 
cancers. While it is impossible to detect definite fallacies in all of these cases, 
especially in that of Dagonet, there are many difiiculties in the way of their 
acceptance. A much greater number of negative results are recorded by 
many observers, as Shattuck and Ballance, Fischel, Sticker, Hemmeter, 
and many others. Herzog found that the nodules first forming invariably 
disappeared if the animal survived long enough and this has been the common 
experience of many pathologists working in this field. The experience of the 
last decade with transplantable tumors, showing how narrowly balanced is 
the nutrition of tumor-cells, renders it extremely unlikely that human tumor- 
cells can proliferate extensively in distantly related animals. The frequency 
with which spontaneous tumors occur in laboratory animals has been greatly 
underestimated by most observers, but has been clearly pointed out by Tyzzer. 

The apparent impossibility of inoculating tumors from one lower animal 
to another of the same species long stood as evidence against their parasitic 
nature. Since the transfer of tumors among lower animals has become 
extensively practised it has become evident that this inoculability does not 
favor the parasitic theory but rather stands strongly against it, since it has been 
shown that the tumor-cells are transplanted and no evidence of a parasitic 
agent has been obtained. It is, of course, conceivable that a parasite is trans- 
ferred with the cells and maintains their tendency to growth, but no evidence of 
such a parasite has thus far been secured, and there is no indication of the 
existence of any such form of parasitism anywhere in the animal or vegetable 
kingdom (Lubarsch). 

Therefore the results of the study of transplantable tumors are every- 
where regarded as a new and serious obstacle in the way of the parasitic 

The development of epithelioma following x-ray burns is a phenomenon 
which upon analysis seems to prove that this tumor develops entirely apart 
from any parasite and arises through slow disturbance of the nutrition and 
mechanical relations of the epithelial cells. 

Clunet has succeeded in producing a malignant sarcoma in a rat by 
repeated exposures to x-ray and, in a case reported by Senger, round-cell 
sarcoma seems to have developed after a lupus cancer treated with .T-rays. 

Chemical agents of great variety have been employed by many observers 
to produce tumors and not without a certain success. 

Emanating from a different point of view, but falling in the same general 
class, are the results of San Felice, Schmidt and Galeotti and PentemaUi, 
previously mentioned, who by means of extracts of yeasts seem to have suc- 
ceeded in rare instances in producing genuine malignant tumors in lower 
animals. Since the great majority of such attempts are failures, the occa- 
sional success may be accepted without in any way involving the conclusion 
that such agents are commonly at work in producing spontaneous tumors in 
man or animals. They seem merely to illustrate the indirect action of 
irritants on predisposed tissues, which is a principle of tumor genesis long 
.since established by clinical and pathological studies in man. 

Likewise the striking observations of Rous that filtered extracts of a 
transplanted and very virulent sarcoma of chickens causes the growth of the 


tumor in a considerable proportion of cases, is probably to be classed as an 
instance of chemical stimulation of cells. 

This observation has the additional interest that the chemical deriva- 
tives of this tumor possess an unusual capacity to excite the neoplastic 
hyperplasia and recalls the well-known observations frequently made in 
man, that normal cells are drawn into the tumor process by the gradual 
diffusion of some influence from the tumor-cells. As an alternate hypothesis 
one may assume that an invisible microorganism passed through the filter, 
but here again the evidence of the existence of such a parasite is defective. 

There are several other observations that chemical agents arising in the 
course of tumor growth possess a notable capacity to excite tumor-like 
growths of tissue cells and it seems probable that in some instances genuine 
neoplasms have been produced by such agents. 

By repeated subcutaneous injections of extracts of human cancer filtered 
through porcelain Mayet claims to have produced cancerous lesions in 
white rats, and Francotte and DeRechter report similar results with cancer 
juice in white mice. Hemmeter by local injections of filtered extract of 
cancer of the stomach from the dog has produced cancerous-like prolifera- 
tion in the walls of experimental gastric ulcers in the dog. 

Theoretical Objections to the Parasitic Theory. — It is often assumed that 
the establishment of a parasitic cause of cancer is surrounded by no more 
difficulties than those which beset the search for the causes of syphilis or 
other infectious diseases. It should be pointed out, however, that there are 
many theoretical objections which would render unusually difficult the 
establishment of any cancer parasite even though supported by strong 
objective data. 

The known infectious diseases display features of incidence, course, 
and anatomical character which sharply separate them from maHgnant 
tumors. The study of the age and sex incidence of cancer presents a body 
of data which reveals a fundamental difference in the etiology of cancer and 
all known infectious diseases. The distribution of cancer is as widespread 
as inflammation in general, indicating that blastomatosis is not a subdepart- 
ment of the reaction of tissues to injury, but a distinct and separate phenome- 
non in the life of the cell. Over against degeneration and inflammation 
stand regeneration and neoplasia. The anatomical and physiological 
characters of malignant tumors differ essentially from those of known infec- 
tious processes. The isolation of the cells of origin of tumors is wholly differ- 
ent from anything recognized in parasitic diseases. The abnormal size of 
nucleus and cell body revealing overnutrition is contrary to the rule in infec- 
tious processes. The types and degrees of metaplasia observed in tumors 
find no parallels in inflammation. The progressive growth of malignant 
tumors reaches a degree which constitutes it a different pathological entity 
from the regenerative process in the healing of wounds and the reaction 
to irritants. Parasitic diseases cause regressive processes in the infected 
tissues and general deterioration of the system by toxic agents, but tumors, 
exert no deleterious action on the system or on the organ involved or even on 
the cells affected except through secondary agencies. The metastases of 
tumors reveal conditions wholly different from any phenomena observed 
with infectious diseases. There is no more impressive illustration of the 
difference between tumors and infectious granulomas, which they most 
nearly resemble, than the comparison of the fate of tumor-cell emboli and of 
emboH from a tuberculous focus. In the former case the tumor-cells grow 
where they lodge, receiving only nutriment from the blood of the part; 
in the latter case the emboHc cells die and the transported baciUi excite an 


inflammatory process in the adjacent tissues. Similarly, transplantable 
tumors survive in the progeny of the transferred cells. 

It is conceivable that an invisible parasite lives with the tumor-cell 
stimulating its growth and nutrition but there is no evidence of any such form 
of parasitism anywhere in the animal kingdom. Parasites occur in mitotic 
cells, especially in club root, but this process is not a tumor (Tubeuf). Where 
parasitism is known it is at the expense, usually, of the Hfe of the cells, while 
degenerative processes in tumors are secondary. 

Weigert and Roux assume that the growth capacities of the cells are 
determined in the ovum and can never be increased except by fertihzation, 
and Ribbert concludes that a tumor-producing parasite is therefore incon- 
ceivable. Yet it must be admitted that the growth of cells is greatly influ- 
enced by the environment and that peculiar external irritants of parasitic 
origin may greatly stimulate the nutrition and growth of cells. It has also 
been assumed that tumors result from release of the restraints of growth, 
but our knowledge of the nature of growth restraints seems hardly sufficient 
to warrant the conclusion that parasites may not be concerned in aboUshing 
these restraints. Biology cannot argue the cancer parasite out of existence 
but it can demand objective data in its support. 

The general facts of the genesis of tumors are strongly against the possi- 
bility of a parasitic origin. 

Tumors arise in some instances from a single ceU (teratoma testis), 
in most cases from a narrowly circumscribed group of cells, and ■ grow chiefly 
or exclusively from their own resources. The gradual inclusion of neigh- 
boring cells in the process is as weD. explained by the diffusion of chemical or 
other influences as by the transfer of a parasite, while there is much evidence 
directly in favor of the former hypothesis. The embryonic nature and 
isolation of the cells of origin of a large proportion of tumors, as embodied in 
Cohnheim's theory, are incompatible with a parasitic origin of such tumors, 
and yet they possess all the quahties of malignant neoplasms which are not 
known to be derived from embryonic cells. The occurrence of highly maUg- 
nant congenital tumors in several members of the same family, as glioma 
of retina, reveals an embryogenic and hereditary disturbance as the essential 
factor. Clinical experience strongly impresses the importance of chemical 
and mechanical irritants and various disturbances of nutrition as the exciting 
causes of tumors. Here stand the numerous Ust of occupational cancers, 
from paraffin, anihn, chimney-soot, the predisposing influence of arsenic 
(Dubreuilh, Hutchinson) and tobacco, »-ray cancers, the influence of chronic 
mechanical irritation, and severe trauma. Likewise cUnical observation and 
anatomical study reveal the fact that cancers which are not of embryogenic 
origin do not arise except after long continuous previous change in the tissue. 
These changes do not favor the estabhshment of any known form of parasit- 
ism but rather suggest disturbances in the nutrition and function of the 
ceUs. Thus BiUroth expressed the conviction that without previous changes 
•in the originating tissue cancer does not exist, and his report of an epithehoma 
arising after many years over most of the area of an extensive scar from a 
burn, well illustrates the basis for his belief. 

Few writers have ventured to suggest that benign tumors can be of 
parasitic origin, and yet occasionally tumors which are otherwise indistinguish- 
able from benign growths, as adenoma of the thyroid and leiomyoma, may 
exhibit all the characters of mahgnancy. 

The developmental history of many tumors exhibits the natural unfolding 
of embryonic potencies. However extensive may be the scope of metaplasia 
in tumors, it cannot cover the facts observed in the field of malignant tera- 


tomas. The teratomas arising from sex cells yield many pure forms of 
benign and malignant tumors on the one hand and highly complex structures 
approaching a parasitic fetus. The metastatic cells from such tumors again 
go on to develop specific structures in lymph-nodes and distant organs. 
Thus in a metastasis of an ovarian teratoma Lubarsch observed brain tissue 
and ependyma in orderly arrangement. Metastases of many tumors are 
distinctly organoid in character maintaining polarity in arrangement and 
functionating as organs. If the above phenomena are of parasitic origin 
then, as Wilms points out, the whole history of embryonic development 
must be conceived as within the scope of parasitism. Hopeless dilemmas 
arise when one attempts to conceive of the necessary properties of the cancer 
parasite. It must pick out minute aberrant groups of embryonic cells in 
protected situations even in the fetus of a healthy mother. Misplaced and 
embryonal tissue invaded by metastases escapes infection by the parasite. 
Shaffer and Lubarsch have described misplaced islands of gastric mucosa 
in the esophagus invaded by epithelioma of the esophagus but showing no 
h)'perplasia. Berent observed a misplaced adrenal rest in the kidney 
unaffected by a metastatic nodule from epithelioma of the jaw. With the 
most actively growing metastases adult normal tissues are not infected but 
as a rule behave passively. In chorioma the parasite invades the fetal but 
spares maternal tissues. In the metastases of melanoma the parasite produces 
extremely rapid proliferation of cells in one case, or it remains dormant for 
many years. 

Contrary to all known forms of parasitism the cancer parasite stimulates 
growth and nutrition of cells and renders them viable after transplantation 
in the same or other animals. It stimulates the absorption of nitrogen by 
the host (Cramer) and fails, except through secondary influences, to exert 
any toxic or deleterious action on the body. 

In view of the considerations thus briefly reviewed it is impossible to 
regard as a valid hypothesis the conception of a specific group of parasites 
living in symbiosis with the cancer-cell and stimulating its growth and 
nutrition. All the facts are reasonably explained by regarding the cancer 
parasite as the cancer-cell. The temporary popularity of the search for a 
specific parasite must be attributed to the undue influence of the germ theory 
of disease which can be effectually combated only by further knowledge 
of the biology of the cell. 

The Scope of the Relation of Parasites to Cancer. — Although the concep- 
tion of a specific cancer parasite living in symbiosis with the cell and stimu- 
lating its growth is without definite foundation and is incompatible with the 
nature of many tumors, there remains a certain field for parasites as etiolog- 
ical factors in tumors. 

In the familiar coccidiosis of the rabbit's liver parasitic ova excite a 
tumor-like proliferation of the bfle ducts. The process is strictly dependent 
upon the presence of the ova and regresses when these are removed and 
hence it is not a true tumor, but it has some of the qualities of a neoplasm and 
these are dependent on a parasitic irritant. 

In Bilharzia infection cancer of the bladder follows in a certain pro- 
portion of cases. Here the disease is true cancer, it is less intimately con- 
nected with the presence of the parasite, but it is nevertheless rather obviously 
a sequel of the infection and reveals a peculiar capacity of this form of 
irritation to produce cancer. 

Lowenstein has described papilloma of the bladder in rats infested with 
Trichodes. Primary carcinoma of the liver in cows suffering from Distomia- 
sis has been referred to the irritation of this parasite (Haaland). In a pap- 


illomatous gastric tumor of pigeons Wasielewski has found a species of 
Dispharagus. Brumpt describes an adenoma of the stomach in monkeys as 
referable to infection by Physaloptera. Febiger has fully demonstrated 
the relation of Spiroptera to a gastric carcinoma in rats. Askanazy attributes 
to a trematode, ApistorcJiis felineus, an etiological relation to certain* hepatic 
carcinomas in man, and similar relations have been claimed for Distomum 
spatulatum and Japonicum by Katsurada and others. 

Many observers have considered the possible relation of trichinosis 
to carcinoma and several have stated that carcinoma is peculiarly frequent 
in chronic trichinosis (Langenbeck, Babes) . I rather frequently find trichinae 
in cancerous tongues. 

Borrel has observed several cases of sarcoma and adenocarcinoma of the 
rat in which a nematode worm was found in the center of the tumor. Regaud 
and Saul have reported similar cases and Bridre found several sarcomas in rats 
growing about cysticerci. Borrel finds acari in the early spontaneous lesions 
of the lymphosarcoma of dogs. In 1 2 early epitheliomas of the face in man 
he found numerous acari inclosed in the tumor-cell nests. In the nipple 
of certain cases of Paget's disease and cancer of the breast he finds acari 
or another animal parasite, Demodex follicidorum. Borrel admits that these 
same parasites occur not infrequently in the healthy skin or nipple where 
they usually produce hypertrophy of the sebaceous glands. These ecto- 
parasites he regards merely as the carriers of the true cancer virus. 

Saul has also drawn attention to the occurrence of worms in mouse 
tumors, and in two human ovarian tumors he found a parasitic mite, Tarsone- 
mus. It has been suggested that the Tarsonemus gained access to the 
material after its removal from the body. 

All of these observations indicate that animal parasites or their deriva- 
tives have a peculiar capacity to excite hyperplasia in the tissues they infect. 
Yet it has not been proven that they are present in any but superficial tumors, 
or that their presence signifies anj^thing more than a secondary invasion. 
Nor is it known that the tumor-like processes in which they are usually found 
have any other significance than the hyperplasia excited by Sudan III which 
regresses upon the removal of the irritant. In lower animals especially, 
animal parasites of the skin are very common while cutaneous tumors are 

Certain manifestations of syphilis have a close though indirect relation 
to tumors. The lingual cancer following leukoplakia may be regarded as 
the result of the disturbed nutrition and relations of the epithelial cells long 
established in the disease and yet the frequency of this form of cancer 
reveals a certain unusual capacity of syphilis to excite neoplastic growth. 
Likewise the syphilitic sarcoma described by Hansemann shows this same 
peculiar influence exerted on mesoblastic cells, the full scope of which is 
possibly not yet recognized. 

Tuberculosis in several of its phases has an indirect relation to certain 
tumors. In a small proportion of cases of lupus, epithelioma develops 
through long disturbance of the nutrition and relation of the epithelial cells 
of the affected area. That there is something specific in the tuberculous 
process is suggested by the greater frequency of epithelioma after lupus 
than with simple chronic eczema. 

In some forms of tuberculous lymphadenitis, tumors of endothelial 
cells develop under conditions strongly suggesting that the neoplasm is a 
sequel of the tuberculous infection. Tuberculosis may cause a widespread 
hyperplasia of the lymphatic tissues of the body with miliary lymphomas 
in many organs. TThe frequent association of lymphatic leukemia with 


tuberculosis has suggested to many that this disease, which has several 
characters of a neoplasm, may be dependent on tuberculous infection or 
intoxication. Hodgkin's granuloma which is either tuberculous or caused 
by a related organism, has been shown in several cases to pass over into a 
sarcomatous process. 

Hence the sequels of both syphilis and tuberculosis have a definite 
relation to tumors, which reveals the microorganisms as indirect cancer 
parasites. In both cases the resulting tumor process seems not to be depend- 
ent on the presence of the parasite but to arise from the natural momentum 
of the disturbance originally excited by the parasite. 

Fig. II. — Growth resulting in 31 days from inoculation in sugar beet of Bacillus lumefaciens 
obtained from poplar gall. (After Smilh.) 

In the lower animals, especially in rats, the momentum of inflammatory 
processes which may lead to tumor growth seems distinctly less dependent 
on the presence of the parasites than in man. 

In plants the observations on crown gall, especially those of Smith, 
point to the conclusion that in the vegetable kingdom progressive hyper- 
plasias originally excited by parasites may be still less dependent upon their 
continued presence than in animals. 

All these considerations encourage the search for new microorganisms 
which may have a special capacity to excite inflammatory processes which 
tend to go on to tumor growth, but they offer no support to the theory of a 
specific cancer parasite living in symbiosis with the cancer-cell and constantly 
stimulating its growth. 

The results in this field show that parasites may be concerned only with 



the inception, of certain tumors but that their influence cannot explain the 
continued autonomous growth of maHgnant neoplasms, wherein Hes the real 
mystery of the cancer process. 

The study of the parasitic relations of cancer suggests the following 
classification of tumors with regard to their possible connection with a para- 
sitic origin: 

I. Tumors of embryonal origin and their derivatives to which Conheim's 
theory appUes, and to which parasites have no relation whatever. 

Fig. 12. — Centers o£ cell proliferation in daisy stem inoculated with Bacillus tumefaciens. 

{After Smith,) 

2. Malignant tumors of exposed surfaces or internal organs with which 
parasites may figure as occasional indirect or direct exciting causes. 

3. Sarcomas, especially those of lymphoid t}^e, which are known to 
be the sequel of infectious diseases or on account of their imperfect neo- 
plastic quahties may probably be regarded as more or less dependent upon 
the presence of parasites. 

4. Tumors of lower animals in which the momentum of a parasitic process 
tends, more than in man, to assume autonomous qualities. 



Among the various fields of the experimental study of cancer the oldest 
and probably the least fruitful is the attempted experimental production 
of tumors. Today it may be said that no one has succeeded in producing 
cancer under conditions that are strictly experimental, although cancer and 
sarcoma have been observed to follow x-ray inflammation in man and ani- 
mals,_ and Febiger has observed the development of cancer on chronic 
gastritis produced by feeding rats with cockroaches infested by a nematode 
worm. Febiger's work may probably be accepted as presenting the con- 
trollable conditions of an experiment, but the proportion of cancers pro- 
duced was small. 

Efforts to produce tumors have passed through several phases suggested 
by theoretical views of the nature of tumors. 

I. Transplantation of adult tissues or cells was early found by Zahn and 
Leopold to result in eventual and usually prompt absorption. Squamous 
epithelium from the combs of fowls Kaufmann found to persist for an un- 
usual period, producing small cysts, and thus duphcating the observed mode 
of origin of traumatic epidermoids. 

Lengemann followed the fate of misplaced cells of liver, kidney, and car- 
tilage. He failed to observe any features of tumor growth, and concluded 
that misplacement alone, hyperemia, or partial degeneration of the cell 
mass, were insufficient to excite tumor growth, while the weakened state 
of the animal had no infiuence on the fate of the cells. In his hands salivary 
gland and thyroid tissue exhibited the greatest proliferation. 

Stilling chose the spleen as the most suitable soil and after transplanting 
fragments of uterine muscle and mucosa in rabbits found in some cases a 
growth of muscle-cells far exceeding their normal limits and resembhng 
fibromyoma, and in other cases he observed the development of large cysts 
lined by proliferating epithelium. 

Nichols in an extensive series of tests found definite proliferative powers 
of adult transplanted cells only in epidermis, uterine epithehum, cartilage, 
and placenta. I. Levin used cells from healing wounds without result. Pass- 
ing interest was excited by Lack's report that multiple carcinoma developed 
in a rabbit's abdomen 14 months after scraping and leaving loose germinal 
epithelium from the ovary. Yet spontaneous carcinomas arise in this 
region in the rabbit (Boycott), and Fraenkel failed in several attempts to 
duplicate Lack's results. 

Lubarsch appears to have encountered a similar coincidence in finding 
a large renal adenocarcinoma in the kidney of a rabbit into whose kidney 
he had some months previously transplanted a fragment of salivary gland. 

Ribbert implanted portions of many organs in the peritoneum, skin, 
anterior chamber of eye, and lymph-nodes, always with eventual atrophy. 
He concluded that fully organized tissue masses might become implanted in 
these new positions and go on to functionate. The transplantation of whole 
organs, or large portions of ovary, thyroid, testis, mamma, has been accom- 



plished by many observers. The transplanted thyroid may reKeve myxe- 
dema (Eiselsberg) ; the transplanted ovary is said to have led to pregnancy 
(Krauer); the testis atrophies without necrosis; and transplanted mammary 
tissue has secreted milk in gestation (Ribbert). 

2. Transplantation of embryonal tissues has been pursued to a formidable 
extent with the general result of showing that embryonal tissues exhibit much 
greater proliferative capacity than adult but still lack an essential factor 
in tumor growth. Loeb, however, found certain adult tissues to grow better 
than the embryonal. 

Leopold concluded that the younger the embryo the better was the growth, 
while Tiesenhausen specified five days as the optimum age of chick embryos. 
Zahn correlated the growth with the extent of vascularization of the trans- 
plant. Cartilage is said to have increased twenty times in bulk in Nichols' 
animals, and Leopold thought he obtained true chondromas. Intermittent 
heating of the organ seemed to increase the growth in the rabbit's hver of 
fragments of periosteum injected by Birch-Hirschfeld and Garten. Preg- 
nancy was found to create a more favorable soil in the work of Askanazy 
and Jentzer in rats, but Shattock found it less favorable for cartilage growing 
in rabbits, while Rous found the most favorable period in mice to be after 

Successive transfers of transplanted tissues at eight day intervals seemed 
to increase the growth in Wilms' experiments with chicks, but an opposite 
result is recorded by Fichera. In most circumstances transplanted embryo- 
nal cells exhibit a certain amount of differentiation as well as of growth but 
in the direction of normal tendencies rather than toward tumor growth. 

Transplantation of complex tissues or portions of embryos, intact or 
comminuted, usually results in abortive teratoid growths. Freund secured 
such growths in 74 per cent, of rats inoculated intraperitoneally with emul- 
sified rat embryos. Such emulsified cells grew for six weeks in rabbits (v. 
Hippel), or eight weeks in chicks (v. Tiesenhausen), btit gradually regressed 
after 6-12 months. Fichera estimated the growth in rats at twenty times 
the original bulk but saw no resemblance to true tumors. Cartilage and 
epiblastic tissues are usually the m.ost active. 

Rous mixed embryo with tumor-tissue, finding that both components 
grew in association but the tumor-cells soon outgrew the normal cells, or 
both failed together, but growing tumor-cells did not stimulate growth in 
the others. 

Traina implanted whole limbs and complex masses of embryonal tissue 
in the ovaries of guinea-pigs and secured striking degrees of proliferation, 
but no tumors. 

The negative results of this extensive department of experimental re- 
search become easily explicable when one compares the obvious crudity of 
the experiments with the minute analysis of pathological data bearing on the 
conditions under which misplaced cells are observed to give origin to tumors. 
Thus among essential conditions for the neoplastic growth of misplaced 
cells are the following: (i) The cell group must not possess the intrinsic growth 
restraints of an organ. Otherwise it will grow or functionate as an organ. 
Evidently there is a nice adjustment between the demands for function and 
the capacity for growth in misplaced cell groups that develop tumors. (2) 
The misplaced tissue should not be completely divorced from its normal 
environment but only partially so. Otherwise it will fail to receive 
sufficient blood-supply for active growth. (3) A considerable period is 
usually necessary for the misplaced cells to adjust themselves to the abnor- 
mal environment. Exceptions to this condition appear probably in some 


traumatic sarcomas. (4) An unusual source of hyperemia is usually neces- 
sary for tumor growth, for this factor facihtates nutrition, and lessens re- 
straints to growth, (s) The misplaced cells must possess some special 
capacity for growth either by virtue of an embryonal character or from the 
retention of proUferative tendencies of more than ordinary degree. Thus 
tumors do not arise indifferently from all portions of an organ but from 
the cells of ducts or from the germ centers of lymph-nodes. (6) The experi- 
mental studies ignore the importance of the prehminary changes which are 
observed to precede the development of tumors from normal or misplaced 
cells. These precancerous stages, the "indifferent vorstadium" of Ribbert, 
are not provided in the experimental transplantation of tissue cells. 

Some of these objections have been met in the efforts to alter the char- 
acter and accelerate the growth of cells before transplantation. G. Lewin 
injected emulsion of a human ovarian carcinoma into the peritoneum of a 
dog. After several weeks the omentum was found studded with miUary 
nodules which seemed to be of inflammatory character. Those were again 
emulsified and injected and the process repeated through five series of dogs, 
the resulting nodules presenting eventually a structure which Lewin regarded 
as sarcomatous. These observations have not yet been verified. 

In somewhat the same manner Stieve produced granulomatous inflam- 
mation about injections of infusorial earth, transplanted this granulomatous 
tissue, and saw more and more atypical prohferation approaching that of 

Influence of Lipoid Solvents on Epithelial Proliferation. — A series of 
experimental studies undertaken from various view-points have indicated 
that ceUs treated or influenced by lipoid solvents tend to exhibit increased 
and atypical proliferation. 

Reinke, noting atypical epithelial growth after the injection of 4 per cent, 
ether into the eye of a salamander, transplanted the proliferating epithe- 
lium into the peritoneum of other salamanders where it continued to grow 
more atypically so as to resemble carcinoma. Askanazy mixed emulsion 
of rat embryos with ether before inoculation and secured larger teratoid 
tumors than by means of untreated embryos, but Freund could not duplicate 
these results. 

B. Fischer introduced an interesting field of study by inoculating under 
pressure a saturated solution of Scharlach R in olive oil under the skin of 
rabbits' ears. The basal epithelium of skin and hair follicles after a few days 
began to proliferate and continued to grow downward inclosing oil globules 
and producing a picture resembling epidermoid carcinoma, but the growth 
ceased when the oil had been absorbed. In the mammary gland of rabbits 
a saturated ethereal solution of the dye produced a squamous epithelial 
growth replacing gland lobules. Since no such effect followed injection of 
simple oils and fats Fischer assumed that the dye contained a specific chemo- 
tactic substance (attraxin) stimulating epithelial growth. Many experi- 
menters verified these observations, offering various explanations of the 
result. The wide variety of materials used in this connection and the general 
results obtained have been conveniently tabulated by Haga. 

Meyer found that when the arteries of the ear were ligated olive oil alone 
produced the results, which he was inclined to refer to chronic inflammation. 
Stoeber and Wacker found certain components of Scharlach R. and Sudan 
III equally effective, and by using 5 per cent, solutions of indol or skatol in rab- 
bit fat they produced very active growth more closely resembling epidermoid 
carcinoma. White produced the usual changes with oleic acid, and Benthin 
used a great variety of materials, finding that the best were Sudan III and 



Scharlach R. These dyes have shown a variable influence upon trans- 
plantable tumors of animals. 

In the mammary glands of rats (I. Levin, White), in the kidneys of 
rabbits (Meyer), and in the prostate, seminal vesicles, liver, and lung of 
rats (Bullock and Rohdenburg), injections of Scharlach R. have produced 
considerable atypical proliferation of inflammatory character, bat much 

Fig. 13. — Atypical epithelial proliferation of epithelium caused by injection of scharlach R. 

(Bullock, Rhodenburg.) 

less marked than in the skin of the ear. Yet Haga succeeded in producing 
markedly atypical bat self -limiting proliferation in the tongue, stomach, 
breast, and liver of rabbits, which in some cases was difficult to distinguish 
from malignant growth. 

Reviewing these results, in general it appears that the property of certain 
lipoid solvents to excite epithelial proliferation has been established but 
no true tumors have been produced (Wacker, Schmincke, Lit.). 

Fig. 14. — Adenomatous growth of the rabbit's stomach, produced by feeding lanolin. 

(After Haga.) 

That the lipoid solvent property is an essential factor in the result, 
although suggested by much collateral evidence, appears uncertain. J. 
Loeb concluded that it was the lipoid solvent properties of the sub- 
stance employed that produced artificial parthenogenesis. The theory of 
solution of a hpoid cell membrane has been pursued by Clowes as the key 
to atypical cell proliferation. Bullock and Rohdenburg reject this theory 



on the ground that all lipoid solvents do not act as does Scharlach R. They 
secured considerable effect from injections of pine tar, indigo, and calomel, 
in oil, and they refer all the effects to cell death and regeneration. Never- 
theless it is clear that certain fat stains exceed all other irritants so far 
tested in their ability to excite epithelial proliferation. That they possess 
somewhat specific properties in this respect is also indicated by their effect- 
ive use to stimulate epithelial growth of ulcerating surfaces. 

After exciting proliferation by means of Sudan III, Lamezan, and Haga, 
have each attempted to transplant the growing cells to other tissues of 
the same animal or to other animals of the same species but without success. 

Fig. is. — Epithelial proliferation of lining cells in ligated duct of salivary gland. (,After 


Chronic Inflamniation in the Experimental Production of Tumors.— 

Many experimenters have employed various forms of chronic inflammation 
in the endeavor to produce tumors. 

Ribbert produced small but typical fibro-epithehal papillomatous 
tumors of the rabbit's lip by repeatedly scraping the regenerating epithelium. 
In many organs he saw atypical growth of mechanically displaced epithehum, 
but in no instance was a true tumor observed. One of his most notable results 
was obtained by ligating the ducts of salivary glands in rabbits which was 
followed by' extensive atypical overgrowth of duct epithelium approaching 
the appearance of duct carcinoma. 

Brosch assumes that preexisting chronic productive inflammation is an 
essential condition in the experimental production of tumors by means of 
chronic irritants. After cauterizing and crushing a portion of the skin 



of the back in white rats he rubbed melted paraffin into the inflarned area 
over a period of 8-12 weeks and secured atypical epithelial proliferation 
closely resembling beginning carcinoma. 

Cazin employed a variety of traumatic influences in the unsuccessful 
effort to produce cancer from normal epitheham. Over a period of 5 
months he rubbed soot into the inflamed skin of a dog without producing 
any atypical proliferation. 

The particular chemical agents- beheved to be responsible for many of 
the forms of cancer observed in the trades, as in workers in paraffin, coal 
tar, soot, tobacco, etc., have been rather extensively employed in the study 
of experimental atypical epithelial proliferation, but in no case has a pro- 
gressive tumor resulted (Haga, Lit.). Special interest attaches to the rather 
notable influence of tobacco extract in exciting epithehal proliferation. 

Fig. 16.- 

-Gastric carcinoma in the rat, caused by infection with Spiroptera. 
{After Febiger.) 

Effect of Extract of Human Ttimors. — Wacker and Schminke injected 
into the rabbit's ear fat extracts from mammary carcinoma, chondroma, and 
gastric carcinoma, and secured moderate epithelial proliferation. I have 
employed the alcoholic extract of a comedo-carcinoma of the breast in 
intramammary injections in the rabbit, securing only an inflammatory reac- 
tion. These experiments were suggested by the impression that the ex- 
tension of comedo-carcinoma in the breast keeps pace with the diffusion of 
fatty material in the obstructed ducts. 

By mixing the expressed juice of carcinoma tissue with Sudan III and 
injecting into the stomach wall of rabbits Haga secured well-marked tume- 
factions of the mucosa with considerable ingrowth of glandular epithehum 
presenting several histological characters of carcinoma. 

Experimental Production of Tumors by the X-ray. — While the frequency 
of development of carcinoma from a;-ray dermatitis places these observa- 



tions virtually in the experimental class, deliberate employment of the x-ray 
for the production of tumors has been reported by Marie, Clunet, and Raulot- 
La Pointe. These authors describe the exact dosage and the intervals of 

Fig. 17. — Position of spiroptera in center of carcinoma of stomach in rat. {After Febiger.) 

treatment followed in a successful effort to produce a malignant tumor in 
the rat. The growth was a spindle-cell sarcoma and proved transplantable 
in other rats. 

Febiger's Experimental Gastric Carci- 
noma in the Rat. — Having discovered the 
remains of a nematode worm in a spontaneous 
gastric carcinoma in a rat, Febiger identified 
this parasite as one occurring in cockroaches 
(Periplaneta Americana, and P. orientalis). 
He designates this parasite as Spiroptera neo- 
plastica. It was later found to infect other 
hosts. He next discovered a source of in- 
fested cockroaches in a Danish sugar factory 
and by feeding a series of rats with the bodies 
of these insects observed the development of 
chronic inflammation, papillomatosis and car- 
cinoma of the stomach. Of 62 rats which 
survived the feeding more than 60 days 12 
developed carcinoma. The interval required 
was as brief as 64 days, and in one case metas- 
tases were observed after 104 days. 

The inflammation and papillomatosis affected a large portion of the 
gastric mucosa and reached such an extensive grade as to fill or even occlude 
the stomach. Structurally the process showed extensive papillary over- 
growth of epithelium with marked keratosis, while the supporting tissue 

Fig. 18. — Spiroptera neoplas- 
tica. {After Febiger.) 



was richly infiltrated with eosinophile leukocytes. From this lesion there 
was progressive invasion of the mucosa and stomach wall by proliferating 
epithelial masses producing the structure of carcinoma. In two animals 
pulmonary metastases were observed in which tumor-cells but no worms 
or ova were found. In the gastric tumor numerous spiropterae with ova 
were commonly present within the proliferating masses. 

This brilliant study presents an experimental production of a well-recog- 
nized form of gastric carcinoma in the rat, demonstrates the dependence 
of the malignant tumor upon the presence of a nematode worm, and reveals 
the gradual assumption by the epithelium of the power of independent 
growth apart from the irritant originally exciting its proliferation. 

Fig. 19. — Structure of experimental carcinoma of stomach of rat, caused by Spiroptera. 

{After Febiger.) 

There appears to be no reason for introducing the idea of a constitu- 
tional dyscrasia in the particular animals yielding the carcinomas, nor for 
assuming the existence of a microorganism carried along with the nematode. 


Between Animals of Different Species.— The early period of attempts 
to transfer cancer from human beings to lower animals, from the time when 
Peyrilhe's servant mercifully terminated the sufferings of the dog into which 
he- had injected cancer emulsion, down to the present date, has practically 
demonstrated the impossibility of this mode of transfer. The numerous 
failures recorded in the literature are doubtless far exceeded by the number 
of unreported experiments. Even using anthropoid apes Metchnikoff and 
Roux and Jobhng had negative results, but it is perhaps not to be assumed 
that in these animals whose blood relationship to man is comparatively 
close, a successful transfer may not prove possible. The significance of 


the few alleged but paradoxical successes has already been considered. 

Between Man and Man. — Cornil reported that an unnamed surgeon 
grafted a fragment of spindle-cell sarcoma of the breast into the other breast. 
At the end of two months the tumor reached the size of an almond and 
proved on section to resemble the original growth. With an adenocarcinoma 
a similar result was obtained but the resulting nodule was not sectioned. 
Fortunately no other chnical record of successful dehberate transplantation 
seems available (cf. Milner). Hahn successfully transplanted a portion of 
skin containing a carcinomatous nodule. 

From the period when vaccine treatment was very widely employed 
I have learned of four cases in which tumors grew from injected unkilled 
cells in the vaccine, some of which were mentioned by Coca. The study of 
spontaneous tumor metastases removes the necessity of experimental 
evidence in this field. 

Between Related Species of Lower Animals.— The early observations 
on transplantable tumors of lower animals indicated that successful grafts 
could be made only on animals of the same species and even then only 
under favorable conditions, but later studies have shown that the suitability 
of the soil is of somewhat wider scope and extends with some tumors to closely 
connected or even distantly related species. Thus Sticker transferred the 
lymphosarcoma of dogs to foxes: v. Dungern carried his hare sarcoma into 
rabbits: Funk claims to have secured successive transplants of rat sarcoma 
in rabbits long fed on emulsified rat sarcoma. Murphy was able to culti- 
vate mouse sarcoma in rats subjected to repeated .T-raying of the spleen 
but as soon as the lymphocytic function of the animals was restored the 
tumors regressed. 

Strauch reports an anomalous result of transplanting Ehrhch's mouse 
carcinoma in rabbit. Three weeks after inoculation he secured in the rabbit 
a well defined tumor which recurred after extirpation. The extirpated tumor 
then failed to grow in the mouse but produced an increasingly malignant 
tumor in other rabbits even giving metastases. Happe succeeded in trans- 
planting rat sarcoma into the lens of a rabbit where it grew out into choroid 
and retina. 

Thus it appears that the viability of the cells of certain tumors of lower 
animals is somewhat greater than once seemed likely but is nevertheless 
sharply restricted to species enjoying a definite blood relationship. There 
seems to be no theoretical difficulty in the way of still further adaptation 
of the cells of one animal to growth in related species. 

Between Lower Animals of the Same Species. — It is highly suggestive 
of the factors determining the results of transplantation of tumors that the 
first success was with the most easily transferable growth, the lymphosarcoma 
of dogs. Yet Novinsky in 1876 had only two positive results in 42 attempts. 
Hanau's transplantation (1889) of epidermoid carcinoma of the vulva in a 
rat was rather more impressive since he dealt with a better known malignant 
tumor and secured extensive metastases. Yet Wehr at the same time inter- 
preted the lymphosarcoma of dogs as a carcinoma, obtained metastases in 
lymph-nodes and spleen in inoculated dogs and noted spontaneous regres- 
sions. In 1890 V. Eiselsberg succeeded in transferring to the mesenteric 
tissues a spindle-cell periosteal sarcoma from a rat. 

The first systematic study of a transplanted tumor and one which revealed 
the possibilities of this method of study was that of Moreau (1891). A 
cylindrical cell carcinoma of a mouse he carried through several generations, 
the tumor gradually losing viability. In the offspring of tumor mice as we 11 
as in pregnant animals it grew more rapidly. Traumatism produced metas- 


tases which were otherwise usually missing. Other animals were insus- 
ceptible. The tumor-tissue when not ulcerated was free from microorganT 
isms. Firket carried a spindle-cell sarcoma through three generations of 
rats, and Vehch lost his periosteal spindle-cell sarcoma only after it had 
passed eight generations, when it failed to survive in a new breed of rats. 

In 1901 and 1902 the general revival of interest in the cancer problem 
gave to the studies of Loeb in America and Jensen in Denmark a special 
interest. Loeb (1900) carried a cystic sarcoma of the rat's thyroid through 
40 generations without observing change of structure or metastases, and 
reported observations on other transplantable rat tumors. Jensen (1902- 
1903) reported the transmission of a mouse carcinoma through 19 generations 
again without metastases. Both these authors convinced themselves that 
the tumor grew from the transplanted cells, but their data on this point were 
not decisive and Loeb assumed the existence of'an agent apart from the cells. 
Jensen found the tumor-cells very susceptible to heat, light, drying and 
antiseptics, but was able to preserve their viability for 18 days at — i°C. 
He further announced that he had seen the regression of large tumors in mice 
receiving injections of antiserum prepared in the rabbit. Borrel strengthened 
the evidence in favor of the carcinomatous nature of the mouse tumors by 
observing numerous metastases in lungs and lymph-nodes. With these and 
other studies shortly following, the modern field of experimental cancer 
research definitely opened. 

Of the permanent results of this new era of experimental research it is 
too early to speak with certainty. In many respects the recent work has 
deepened our insight into the conditions of growth of tumors, especially the 
narrow range of their viability and at the same time their apparently un- 
limited capacity for growth in suitable environment. Particularly fortunate 
are the conditions provided for the detection of any immunity factors that 
may be concerned and yet progress in this field has perhaps been more 
limited than in others. It is apparent that many phenomena encountered 
in the study of transplanted tumors and hastily interpreted as new observa- 
tions had long before been fully recognized in spontaneous human tumors 
by those familiar with this field. Many observations on change of structure, 
the apparent discrepancy between histological appearance and biological 
behavior, the mechanism of tumor regression, the conditions governing 
metastasis, the elements contributing to cachexia, have proven httle more 
than welcome reappearance, under experimental conditions, of facts and 
principles previously established in the human subject. The study of 
transplantable tumors has shed no light on the histogenesis or mode of 
origin of neoplasms. In the field of practical therapeutics it has opened 
up many new trails which unfortunately have led mostly astray. Chief 
among the less important results is the further evidence supplied by the 
extension of the work to various lower animals that different tumors are 
often quite different clinical and pathological entities, and that observations 
on one ma}- not safely be apphed to another. 

The problems arising in the study of transplanted tumors may now be 
considered in brief detail. For a full discussion of this subject cf. Woglom. 

Problems of Transplantable Tumors. — Nature and Origin of the Tumors. — 
Many doubts were first expressed regarding the malignancy and even of 
the neoplastic nature of the transplantable tumors. Yet it has been abun- 
dantly shown that the chief tumors in mice often exhibit infiltrative growth 
and metastases. Bashford, Murray, and Cramer described in detail local 
infiltration and metastases in the Jensen tumor, and Apolant showed that 
this type of growth developed when the tumor met dense resisting tissues- 


Murray found pulmonary metastases in about 50 per cent, of his spon- 
taneous tumors but they were often microscopic. The significance of the 
morphology of these tumors must also be regarded as unequivocal. When 
the whole list of mouse tumors now described is considered, it appears that 
this animal is subject to a considerable variety of tumors comparable to those 
of man and a standard of tumor morphology for this animal has been 

The origin of the chief mouse tumor has been shown by Bashford and 
Murray and Apolant to be from widely scattered mammary tissue. Many 
other mouse tumors have been described and traced to their probable origin, 
as epidermoid carcinoma of the mouth (Borrel, Haaland), adenocarcinoma 
of intestine, carcinoma of stomach, adenoma of liver, spindle-cell sarcoma of 
renal region (Bashford, Murray, and Cramer), chondroma (Ehrlich), adeno- 
carcinoma of kidney, adenocarcinoma of ovary, melanoma, myoma uteri, 
adenocarcinoma of sebaceous glands, and many others (Haaland, Tyzzer). 
Apolant thought that 95 per cent, of mouse tumors arise in the mamma, 
but in Tyzzer's series of 70 tumors of mice 74 per cent, arose in the lung. 

In the rat the series of observed tumors is almost equally extensive. 
Loeb's tumor was a spindle-cell growth apparently sarcomatous arising in the 
thyroid. Other sarcomas are described by Herzog and by Jensen. Stohr 
reports a fibro-epithelioma of the tongue, Flexner and Jobling an adeno- 
carcinoma of the seminal vesicle; and Lewin and Michaelis a mammary 
carcinoma; Gaylord, a spindle-cell sarcoma. Many of these tumors have 
been extensively transplanted. 

In the dog, mammary carcinoma is a common tumor which presents 
many peculiarities but all the essential features of malignancy. It has been 
transplanted in the Loomis Laboratory by Teague into animals weakened by 
distemper. The infectious lymphosarcoma of dogs has been variously inter- 
preted. It is a large round or rather polyhedral cell-growth probably arising 
from the reticulum cells of the lymph-node. Beebe and the writer have 
shown in bacteria-free transplants that it arises from the transplanted cells. 
Whatever may be the conditions surrounding the inception of this tumor, 
when fully developed it presents all essential features of a malignant neo- 
plasm. Many efforts to cultivate bacteria have failed, and spirochetae 
occur only on ulcerating surfaces. The conditions permitting transplan- 
tation are similar to those observed with other tumors. 

In the fowl lymphocytomas with and without leukemia are comparatively 
common. In Wernicke's summary of the recorded tumors of fowl it is 
seen that a considerable variety of newgrowths has been observed. Ehren- 
reich found seven tumors of carcinomatous or sarcomatous type among 1000 
old chickens. An attempt to transplant one of the carcinomas failed. Pick 
observed an epidermoid carcinoma of the mouth; Tyzzer reports myxosar- 
coma of the thigh and leiomyoma of the mesentery. Fujinami and Inamoto 
have described a myxosarcoma which was readily transplantable and pro- 
duced numerous metastases in the inoculated fowl. 

Rous has described three chicken sarcomas, spindle-cell, and osteochon- 
dromatous, all of which are transplantable. The juice from the spindle-ceI1 
sarcoma strained through a Berkefeld filter (No. 5, medium) as well as the 
dried and powdered tumor-tissue reproduced the growth. The effective 
agent remained in the dried tissue for seven months but was destroyed by a 
temperature of S5°C., and would not pass a Chamberland bougie (F. i). 
It was very susceptible to antiseptics. Quite remarkable was the increase 
in viability of this tumor, for while it was at first transferred with great diffi- 



culty it eventually gave loo per cent, of successes in Plymouth Rocks. It 
could not be transplanted to other fowl. 

Fig. 20. — Structure o£ nodule of lympliosarconia of dog three days after transplantation. 
Note line of separation between tumor-cells and host's tissues. 

Fig. 21. — Section of a pulmonary metastasis of Rous' chicken sarcoma. 

In animals inoculated with the juice secondary tumors very often devel- 
oped at points of trauma, but always much more slowly than in animals 
inoculated with cells. Filtrates from the spindle-cell tumor and from the 


chondroma reproduced only the original tumor. It thus appears that in 
Rous' sarcoma a tumor-producing agent can be separated from the tumor- 
cells by coarse filtration. The nature of this agent has not been determined. 

It seems probable that the action of this agent may be compared to 
that of substances diffusing from certain human tumors into surrounding 
tissues and causing peculiar collateral hyperplasia and even the gradual 
transformation of normal into tumor-cells. It may also be pointed out 
that various chemical agents have been shown to be specially adapted to the 
excitation of atypical overgrowth of cells. Another factor of undetermined 
value may be a peculiar quality in the reaction of chicken tissues to irritants. 
The principles of avian pathology are not as yet fully expounded. It is of 
little moment to inquire whether or not the Rous sarcoma is a true tumor. 
It is obviously not a simple inflammation but enjoys several essential features 
of a neoplasm. At present it must be regarded as a disease sui generis 
without exact parallel in other animal species. 

In rabbits a considerable number of tumors have been described but 
none appears to have been extensively transplanted. In the hare v. Dungern 
and Coca discovered a slowly growing spindle-cell sarcoma in the skin of the 
ear. It contained spindle-cells mingled with large polyhedral eosinophile 
ceUs and somewhat resembled a granuloma. It proved readily transplantable 
in rabbits, and from the serum reactions of the inoculated animals the 
authors concluded that the tumor-cells even in late generations retained the 
properties of hare protein. The significance of these serum reactions is, 
however, open to doubt. 

Fate of the Transplanted Tumor. — The demonstration of the survival of 
the tumor-cells in the transplanted tumor has been a highly important 
result of experimental studies, since it establishes an essential distinction 
between neoplasms and infectious processes. This principle had, of course, 
been previously demonstrated by observations on metastatic processes in 
human subjects but the experimental evidence has been more decisive. 
Observations at intervals upon the implanted tumor grafts have shown for 
the tumors of Jensen, Loeb, Ehrhch, the lymphosarcoma of dogs, and Rous' 
chicken sarcoma, that while the central portions of the mass suffer simple 
necrosis, the peripheral cells stimulate the formation of new blood-vessels 
and the multiphcation of fibroblasts, and with this support they multiply 
and produce the new tumor. This reaction has been attributed to somewhat 
specific formative powers acting on fibroblasts or angioblasts and producing 
chiefly connective tissue or granulation tissue about the tumor graft (Ehrlich, 
Gierke). A leukocytic or a lymphocytic reaction is also observed, especially 
when there is much necrosis in the graft, or when it is infected. Much 
exudative inflammation often interferes with implantation. V. Dungern 
and Coca describe a phagocytic absorption of tumor-cells in the hare by 
means of large mononuclear leukocytes, which interfere with implantation. 
The stroma of the transplanted tissue usually undergoes hyahne transfor- 
mation and absorption but in some cases it may be preserved for a certain 
period. Thus Beebe and the writer, in the lymphosarcoma of dogs, found 
on the third day arterioles in the graft filled with circulating blood. Pro- 
gressive proliferation of the transplanted stroma has not been satisfactorily 

The time required for the appearance of proliferation varies greatly 
with different tumors. It may be visible in 24-48 hours. A palpable tumor 
seldom appears before 10 days, usually it may be detected in two or three 
weeks, but occasionally is delayed for months (Bridre, Stohr). 


Adaptation. — Great variation in the proportions of successful transplants 
has nearly constantly appeared. Nearly all observers found difficulty in 
securing the first graft, while subsequent generations have usually followed, 
much more readily, so that some tumors, as the Buffalo rat sarcoma, have 
eventually yielded loo per cent, of "takes." Careful analysis of the conditions 
of successful transplantation has been necessary in order to determine whether 
the differences were due to an increase in the growth capacity of the cells 
or to some pecuhar adaptation to the experience of transplantation. The 
former possibility led to the employment by Ehrlich of the conception of 
increased virulence. 

It was soon shown that the chief factors resided in the transplanted 
cells, since within the range of animals commonly employed the increase in 
number of successful grafts and in their rate of growth was invariable. Pre- 
liminary heating to 39-4i°C. or exposure to mercuric chloride or iodide, 
yielded a larger percentage of takes in the experience of Clowes and Baeslack. 
Heat beyond 4S°C. or a few minutes exposure to 2o°C. kills most tumor-cells 
(Jensen, Loeb). Yet Ehrlich transplanted a carcinoma kept for two years 
at — io°C. although the tumor soon ceased to grow. Liquid air at a tempera- 
ture of — i9S°C. failed to kill all cells in the experiments of Moore and Walker, 
while Gaylord and Haaland eliminated the carcinomatous elements from a 
mixed carcinosarcoma by heating to 44°C. for 30 minutes. With the excep- 
tion of Rous' chicken tumors it has always been found necessary to transplant 
intact cells, while extensive comminution or drying abolished growth. Most 
tumor-cells are comparatively resistant to ordinary antiseptics. 

Spontaneous fluctuation in growth energy of the implanted tumor has 
been assumed in order to account for variations in the results of implan- 
tation. As stated by Bashford, Murray and Cramer, "considerable varia- 
tions in success attend the inoculations of one and the same tumor at different 
times. Thus one series of inoculations may give a small percentage of slow- 
growing tumors which at a subsequent period may begin to grow rapidly 
or, on transplantation, while still growing slowl)', give a high percentage 
of quickly growing tumors." From a large series of inoculation experiments 
taking into account the percentage of "takes" and the rate of growth, these 
authors constructed carves showing rhythmic variations in growth energy. 

It is thus interesting to note that in this experimental field conclusions 
may be drawn which support clinical observations frequently made in the 
human subject, that the rate of growth of many tumors varies at different 
periods of their course. Whether the two sets of observations are exactly 
parallel may be doubted. It may further be concluded from clinical and 
pathological study that the rate of growth of different portions of the same 
tumor varies enormously, the active portions being usually peripheral, that 
one tumor may be active while another neighboring mass is quiescent, that 
superficial tumors may regress when internal metastases are active, and that 
in general the rate of growth of tumors is subject to fluctuations from a great 
variety of factors which are chiefly dependent on the uncertain nutrition 
of both tumor and host. Under these complex conditions it is extremely 
difficult to determine whether anything properly designated as increased 
virulence or adaptation of tumor-cells has so far been demonstrated in the 
experimental field. On the contrary it would appear that successive trans- 
plantation involves a process of sifting out the more viable cells and providing 
for them a highly favorable soil. 

Uniformity in technic of transplantation is an obvious essential 'in a 
comparison of results. Emulsions of cells in salt solution, or thin slices of 
tissue, are chiefly employed and in either case is has been found that an 


optimum quantity exists, viz., i to 2 cm. of emalsion and i or 2 cubic 
millimeters of tissue. Larger amounts may yield initially larger tumors but 
in many cases subsequent standstill or regression are observed. 

With large doses it has been assumed that partial immunity resulted from 
absorption of much of the tumor-tissue. 

The site chosen for inoculation is ordinarily the subcutaneous tissue, 
into which the tumor material is injected or implanted. The internal 
organs, especially the spleen, have proven less suitable (Goldmann). Uhlen- 
huth secured grafts by rubbing an incised wound with the Jensen rat sarcoma, 
and Beebe and the writer transferred the dog sarcoma in this same manner 
which is apparently the natural mode of transfer of this disease during 

Influence of Variations in the Soil. — That slight variations in the tissues 
of the host greatly influenced the results of transplantation was very early 
noted. Not only was the same species of animal required but the results 
were more successful in animals of exactly the same color and antecedents. 
Brown and wild strains differed in susceptibility from white and domesticated 
families. Rous found it impossible to transfer his chicken sarcoma at first 
to any but Plymouth rocks, preferably of the same brood. Haaland found 
it impossible to propagate on Danish mice fed on grain a Berlin tumor coming 
from mice fed on milk. Many others have had difficulty in transferring 
propagable tumors from the animals of one country to those of another. 

The infl-uence of diet has since been shown to be highly important. Beebe 
and Van Alstyne were able to render rats highly refractory to the Buffalo 
sarcoma by a previous course of carbohydrate-free diet. They have also 
shown that the course of this tumor is distinctly retarded by carbohydrate- 
free diet, and accelerated by butyrates among the fats. Sweet also has 
observed increased resistance to implantation in animals on a carbohydrate- 
free diet. That these rules may not hold for all tumors is highly probable 
and specifically shown by Woglom's negative results in carcinoma. Benedict 
brought about the complete regression of large sarcomas in rats rendered 
diabetic by phloridzin. 

Simple restriction of food has been shown by Moreschi to retard the 
growth of implanted tumors. With the Flexner-Jobling rat carcinoma Rous, 
however, failed to irifluence the growth by restricted diet. 

By substituting Kme substances for the usual diet Sweet, Corson and 
Saxon rendered rats markedly insusceptible to implanted sarcoma, and 
when a graft succeeded in such animals its growth was accelerated by return 
to normal diet. Castration also favored growth. Centanni reports that a 
diet of wheat bread and corn but without green food renders mice highly 
insusceptible to tumor implantation. Oser and others find that tumors 
grow better in splenectomized animals. 

Spontaneous absorption of implanted tumors was first recorded by Loeb 
and its comparative frequency and the increased resistance to reinoculation 
were emphasized especially by Clowes and Baeslack. About 23 per cent, of 
tumors derived by Gaylord and Clowes from the Jensen tumor were found 
to regress, and Bashford reported as high as 50 per cent. Most of the regress- 
ing tumors have never reached a large size but some very large growths appear 
to reach a critical point beyond which nutrition fails and complete absorption 
follows. It is a notable fact that during absorption the animals do not 
suffer from intoxication and that recovery may be complete. The histolog- 
ical changes in regressing tumors vary widely. A simple necrosis is often 
the primary event followed by absorption and fibrosis, or a progressive 
fibrosis may overtake the tissue and lead to gradual atrophy of tumor-cells. 


Phagocytic processes by mononuclear cells are frequently observed (Gaylord, 
Clowes, Bashford). 

Structural Changes in Propagated Tumors. — While many tumors maintain a 
uniform structure through many generations, some carcinomas have exhibited 
notable changes in structure which are difficult to interpret. As a rule 
these changes duplicate those observed in the human subject, and on the whole 
they appear to be less violent than are frequently observed in human material. 
Thus Bashford's study of 85 propagable tumors, 35 of which had been in 
cultivation for 3 years, revealed the usual series of minor changes in the 
size and structure of alveoli, appearance and disappearance of various forms 
of epithelial metaplasia, formation of bone and cartilage in sarcomas and 
variations in vascularity. Metastases were not always more atjrpical than 
the original growth. Apolant was disposed to attribute the reversion of a 
sohd carcinoma to an adenoma to increased resistance in partly immunized 
animals, a conclusion from which Murray dissents. 

Considering the variation observed in propagable tumors in comparison 
with the structural changes in recurrent human tumors the entire scope 
appears very much more restricted in the lower animals than in man. 

The most important change credited to the propagable tumors is the 
sudden assumption of malignant neoplastic properties by the previously 
normal stroma. This change was first observed by Ehrlich and Apolant 
in the tenth generation of an adenocarcinoma of the mouse. A reverse trans- 
formation, viz., the. disappearance of spindle-cells from an adenocarcinoma 
is also recorded by Apolant, but in this instance it was concluded that the 
spindle-cells were merely altered epithelium. 

Many later observations of the development of sarcoma in carcinoma have 
been recorded in both rats and mice (Loeb, Lewin, Bashford, Russell). 

The change appears to occur rather suddenly after the 8-ioth or later 
generations. RusseU fixed the period at which the change usually appears 
at the S5th to 60th day of propagation and states that it is independent of 
the number of transfers. When once established it is usually progressive 
and in certain cases the spindle-cell tissue has eliminated the other element, 
or both structures persist together, or separate strains of sarcoma and carci- 
noma are obtained. The rate of growth of the mixed tumors appears in- 
creased rather than diminished. In metastases either sarcoma or carcinoma 
were encountered, but the sarcomatous element when present was more 
prominent in metastases (.Haaland, Clunet). 

Several theories have been advanced in explanation of the sarcomatous 
transformation of mouse carcinoma. 

It was at first considered possible that the growths were originally mixed, 
but this theory was abandoned for lack of evidence. 

Most observers have concluded that the sarcoma represents a neoplastic 
transformation of the stroma of the host caused by a stimulation of stroma- 
cells by the epithelium. The exact nature of this stimulus is not clearly 
defined but it has been conceived as a sort of growth stimulus possibly 
carried by chemical derivatives of the epithelium. 

Repeated transplantations are not necessary. Both Haaland and Russell 
describe in detail the appearance of foci of overcellular stroma located in the 
center of carcinoma nodules, the increase of mitotic figures in the stroma- 
cells, the survival of these altered stroma-cells in grafts and their rapid in- 
crease until they overgrow the epitheUal elements. Haaland describes 
peculiar halos of large cells surrounding epithelial groups, and intermediate 
between epithehum and stroma-cells. Others have failed to trace the 
spindle-cells to either stroma or epithelium. 


The third possibiHty, which seems to the writer to have been too hastily 
dismissed, is that the spindle-cells are altered epithelium. Such transfor- 
mations are relatively common in the early or advanced stages of many 
human tumors and especially in recurrences after operation, and this fact 
establishes a probability that a similar change in mouse tumors has a similar 
significance. On the other hand there is no parallel in human pathology 
for the development of sarcoma in stroma supporting carcinoma, so that 
observations in mice would introduce a new principle into oncology. In 
man spindle-cell transformation usually goes with increased growth and 
mahgnancy and such is the case in mice. It is difficult to conceive how an 
original carcinoma can be made to yield all its growth energy to normal 
stroma-cells and completely retire from the field, yet this anomalous result 
appears to be reached in the pure sarcomas developing from mouse carcinoma. 
The crucial evidence is that presented by Haaland and Russell of gradual 
transformation^ of stroma into sarcoma. Yet the interpretation of transi- 
tional pictures is notoriously hazardous, and few observers h.ave been able to 
convince themselves that it actually occurs. Orth thought that Lewin's sar- 

r .,- ,^. 

.■• ■• <it,"',n-.:-i' 'aC» -> fl " ;-». •,■■■- -"-trx I, ' I* • -iass .iij-jf'^.s. ' J*' ^'^■: 

' >■■■■ • - ; - -'•■■<-■■ 'yic .-r • « "^ -v^^y ■^>•24,s%;^.: ,f,« 

Fig. 22. — Halos of pale cells surrounding epithelial cell groups, in the sarcomatous trans- 
formation of carcinoma in the rat. {After Haaland.) 

coma-cells represented granulation tissue. I have examined, through the 
kindness of Woglom, several cases purporting to show the sarcomatous trans- 
formation of stroma-cells, but have been forced to draw from these sections 
the opposite conclusion, viz., that the spindle-cells were derivatives of epi- 
thelium. Since such an interpretation is at least admissible, it may be 
urged that further evidence is required before the sarcomatous transformation 
of mouse carcinoma can be accepted as proven. 

Resistance. — In the elucidation of the factors concerned in natural and 
acquired resistance to tumor implantation the experimental studies have 
opened up a new and highly fruitful field. The results apply strictly to the 
artificial conditions in normal animals receiving tumor grafts, but not them- 
selves developing spontaneous tumors. It has repeatedly been shown that 
animals resisting implantation may often develop spontaneous tumors 
(Bashford, Thorel, Clunet). Likewise very young animals which do not 
develop cancer may prove very favorable soil for implantation. This prin- 
ciple reappears in the rapid growth of many carcinomas occurring in human 


subjects of youthful age. The problem of the continued growth of a tumor is 
therefore quite different from that of its inception. It follows also that 
an animal may be susceptible to one tumor while insusceptible to another, 
and it is equally clear that the body tissues of animals in general are very 
favorable soil for the growth of lawless cells which must therefore be re- 
strained by purely local factors. 

Natural Resistance. — Age has been shown to have no definite influence 
on the susceptibihty to certain tumor grafts since both young and old animals 
have proven suitable for experimental studies. Most observers have pre- 
ferred stock from 4 to 8 weeks of age (Haaland, Albrecht, Hecht). 

Racial differences were beheved to account for the capricious behavior 
of many tumors when transferred from animals of one country to those of 
another. Yet it has appeared that, many indications of resistance seemingly 
referable to racial differences were the result of altered diet, and Gierke 
strongly supported this view. In many instances strange animals at first 
resistant soon became susceptible when housed under the same conditions as 
the animals furnishing the graft (Stohr, Cuenot, Mercier). For absolutely 
comparable results the use of animals of one litter or parentage is probably 

Good health appears to favor the growth of tumor grafts and poor con- 
dition to retard it. Yet Teague succeeded in transplanting mammary carci- 
noma of dogs only in animals weakened by distemper. Observations on the 
effect of pregnancy are conflicting, Moreau and Herzog finding acceleration, 
while Haaland, Bridre and Cuenot saw retardation of growth in gestation. 

Acquired Resistance. — The first observations concerning acquired resist- 
ance were made by Gaylord, Clowes, and Baeslack who noted frequent 
spontaneous regression of the Jensen rat tumor and also the insasceptibihty 
of these recovered animals to reinoculation. Similar results were soon re- 
ported by many observers with various tumors and the question arose 
whether the diminishing proportion of successful takes in reinoculated 
animals was due to a gradual sifting of naturally immune animals or to active 
immunity caused by repeated absorption of the tumor grafts. The latter 
theory was strongly supported by Bridre who found that by inoculating very 
large doses his resisting animals rapidly became quite immune. Michaelis, 
Borrel and Ehrhch verified these results. It next became apparent that 
the resistance thus established while most pronounced against the tumor 
employed extended also but to a lesser degree to other tumors. Ehrlich 
was inclined to believe that the resistance was practically universal (pan- 

That the actual temporary growth of tumor-cells was necessary to 
excite the resistance was strongly suggested by its failure to appear in ani- 
mals treated by tumor-cells killed by heat or mercuric chloride (Clowes); 
or killed by chloroform (Michaelis); or treated with the filtrate of tumor 
emulsion (Bridre); or by cells injured by grinding and freezing (Haaland). 
Whether the principle has been fully established is perhaps doubtful. 

Normal . tissue cells are almost equally effective in exciting resistance, 
as first shown by Bashford using red corpuscles, and later by Schone, using 
liver, testis, or emulsion of mouse embryo; by Borrel and Bridre, with spleen, 
liver, or testis; by Higuchi with blood-free placenta. With lactating mouse 
breast Moreschi secured variable results. Both Russell and Woglom 
produced the resistant state by emulsions of embryo skin and Woglom 
ascertained that the immunity reached a maximum 10 days after inoculation, 
remained high until the twenty-fourth day, and vanished about the seventy- 
fifth day. 


Devitalization of the normal tissue cells generally robbed them of their 
immunizing power, while Woglom showed that injection of the animal's 
own tissues were ineffective. 

When these experiments were transferred to tumor-bearing animals the 
results became less definite, although Russell considerably reduced the 
proportion of successful secondary inoculations when the reimplantation 
was preceded by an injection of tumor or embryo-emulsion. The practical 
test of these methods in animals with spontaneous tumors proved entirely 
fruitless in the hands of Haaland who failed by this form of vaccination to 
influence the growth, prevent post-operative recurrence, or diminsh metas- 
tases from spontaneous tumors. Likewise in the human subject, vaccination 
therapy employed on a vast scale, while attended by some paradoxical 
successes, has failed to establish its usefulness (Coca). 

Passive Resistance. — Attempts to convey actively acquired resistance 
to other animals has not been distinctly successful. The use of the serum 
of resistant animals has usually failed to render new animals passively 
resistant to inoculation (Michaelis, Ehrlich, Haaland). Repeated reinocula- 
tions of immune rats by Uhlenhuth failed to render their serum effective in 
preventing successful inoculation in fresh animals receiving first serum and 
then tumor, or serum and tumor together. Bridre prepared in sheep and fowl 
an antiserum against mouse cancer which did not retard and even appeared 
to accelerate tumor growth. This method has been extensively employed in 
human therapy using as antigen either the whole tumor or various protein 
fractions, with uncertain results (Vidal, Beebe, Berkeley). 

Nature of the Factors in Resistance. — The blood serum having failed to 
yield proof of possession of the immunity factors it became evident that 
resistance to tumor implantation must be of more complex nature than that 
■excited by bacteria. In this emergency Ehrlich pronounced his theory of 
athrepsia to account for the failure of tumor grafts to grow in resistant 
animals. Having found that mouse carcinoma after surviving in rat tissues 
for a week may then be successfully restored to active growth by returning 
the graft to the mouse, he succeeded in conducting a series of zig-zag transfers 
from mouse to rat to mouse, etc., without permanent injury to the tumor. 
From this data he argued (i) that there were no cytolytic agents in the rat, 
otherwise the tumor-cells would have been injured, and (2) that the mouse 
•graft failed to grow in the rat after a few days sojourn because it soon ex- 
hausted a specific nutritive substance, which was not the nutritive molecule 
itself but acted as a stimulus to nutrition. Replacements in the mouse 
provided new supplies of the x-substance. This theory abandoned the idea 
of antagonistic substances in the resistant animal and substituted the con- 
ception of an inert lack of nutriment as the factor in resistance. The theory 
was possibly a correct interpretation of the experiments cited but failed to 
meet the conditions in animals of the same species. Here two valid objec- 
tions were soon raised by many authors (Uhlenhuth, Borrel, Hertwig and 
Poll, Bashford). It has been shown (i) that a tumor-bearing animal may 
usually be reinoculated, both tumors advancing or regressing together, the 
■one not interfering with the nutrition of the other; and (2) that when a 
second inoculation fails it is usually due to an immunity excited by the partial 
absorption of the original graft. 

While thus failing to account for actively produced resistance Ehrlich's 
athreptic theory has served to emphasize the probable importance of the 
nutritive qualities of the soil as opposed to antagonistic substances in the 
blood of resistant animals. 

A further illustration of this principle may be found in the experiment of 



Beebe and Crile who regularly caused the disappearance of the lympho'sar- 
coma of dogs by exsanguinating the animals and immediately transfusing them 
with the blood of resistant dogs. Here it may be assumed that an antagon- 
istic substance was carried over in the blood, but it is simpler and equally 

Fig. 23. — Cultivation of mouse carcinoma-cells in blood-plasma. (Lambert and Hanes.) 

reasonable to suppose that the transfused blood was better adapted to the 
normal than to the tumor-tissue, and that its presence turned the balance of 
nutrition in favor of the normal tissues and against the tumor which then 

Fig. 24. — Rat sarcoma, growing in pigeon plasma. (Lambert, Hanes, Crocker Reports, 


s^uffered slow 'atrophy and absorption. Rather specific evidence against the 
purely athreptic nature of tumor immunity has appeared in the successful 
cultivation of rat ,and mouse tumor-cells in various ahen plasmas (Lambert, 
Hanes). ' 


Cultivation of tumor-cells in vitro. The conditions requisite for the prolonged survival 
or actual multiplication of tissue and tumor-cells in vitro seem to be a fluid medium per- 
mitting of nutrition and diffusion, fixed points of attachment for the cells, and a suitable 
temperature. Nutrient fluids exactly homogeneous with those to which the cell has been 
adapted are not necessary. These conditions were partially provided by Loeb who saw 
slight spreading of epithelial cells over the surface and into clefts of an agar medium. 

The fixed points of support were missing in the circulating whole blood used by Beebe 
and the writer, and we secured only prolonged survival of dog tumor-cells. An adap- 
tation of Harrison's technic by Burrowes proved successful in his hands and Carrel's in 
permitting definite multiplication of tissue cells in the fibrin meshes of coagulated plasma, 
and this method had been extensively used. Epithelial cells grow out in coherent sheets 
from the parent mass, while mesoblastic cells tend to grow out in long sprouts or become 
widely scattered. The importance of the results of experimental cultivation of tissue 
and tumor-cells in vitro has not been as great as might have been anticipated, but it has 
been shown that these cells are capable of living in much more varied media than once 
seemed likely, and that the natural immune substances supposed to exist in the serum of 
refractory animals has little influence on the growth. Tumor-cells as a rule seem to be less 
capable of survival or growth than corresponding normal tissue cells, and only after some 
difficulty has Burrowes succeeded in cultivating certain human tumor-cells. 

Hyper susceptibility. — Various observers have reported phenomena sug- 
gesting hypersusceptibihty to tumor growth in animals receiving grafts, 
emulsions, or saUne extracts of tumors, but the data are insufficient to 
establish the existence of a specific hypersensitization (Woglom, Lit.). 

The mechanism by which the resistant animal aborts the tumor graft 
offers suggestive evidence regarding the nature of the immunity factors. 
In resistant animals the unsuccessful graft fails to excite the specific stroma 
reaction and remains unprovided with the necessary channels of nutriment. 
Russell examining grafts in refractory animals at intervals found that the 
surviving cells continued to proliferate for 7-10 days but instead of exciting 
an active growth of fibroblasts they became sharply separated from the 
host's tissues. He concluded that there must be present in resisting animals 
something which inhibits the chemotactic influence of the cancer-cell upon 
the connective tissues of the host. Inflammatory reaction has also been 
shown to be responsible for the failure of grafting in resistant animals. 
V. Dungern and Coca, studying the inoculation site in immune animals, and 
in regressing sarcoma in the hare, observed that the necrosing graft gathered 
large numbers of macrophages in vessels and tissue spaces, as weU as many 
lymphocytes and plasma-cells, which attacked and absorbed the tumor- 
cells. Since this reaction failed in susceptible animals they interpreted it 
as a phenomenon of hypersusceptibihty or allergic, provided by the previous 
immunizing dose. Burgess examining grafts in non-susceptible Japanese 
mice observed at first the formation of the usual vascular stroma but after 
a week an inflammatory reaction ensued yielding polynuclears, eosinophiles, 
lymphocytes and plasma-cells, which shut off the blood-supply and induced 
fibrosis. Da Fano also noted a rich exudate of lymphocytes and plasma- 
cells about the inoculated tumor in immunizing animals. In splenectomized 
animals Apolant finds that very slight immunity to tumors can be excited, 
and Baeslack observed marked fall in the circulating lymphocytes m sus- 
ceptible animals with growing tumors, and a steady increase in resistmg 
subjects with regressing tumors. Murphy finds that the chicken embryo 
possesses no means of preventing growth of implanted rat tumor untfl the 
chick hatches and develops spleen and lymphocytes, but if the embryo receives 
a fragment of adult chicken spleen or bone-marrow along with the tumor 
graft the latter fails to grow. After subjecting the rat spleen to the a,--ray 
he found that the animal supported the growth of alien tumor-tissue for a 
considerable period and until the spleen recovered its lymphocytogemc 


All these observations lend important support to the view, long since 
established in human pathology, that the lymphocyte is an important 
agent of defense of the organism against tumor growth. 

Some light but no complete elucidation of the nature of tumor immunity 
has been furnished by collateral studies on tissue immunity. 

Antibody production has been demonstrated against various tissue 
cells injected into foreign species; as tracheal epithelium (V. Dungern); 
or spermatozoa (Metalnikoff). These observations may justify the use of 
antisera prepared in lower animals against human carcinoma tissue or its 
derivatives. ' 

Antibody production has been demonstrated against an animal's own 
tissues, injected either into the same individual or into other animals of the 
same species. Thus Adler prepared an autospermatoxin in guinea-pigs and 
Halpern showed, by complement deviation, the development of antibodies 
against an animal's own kidney, pancreas and spleen. These observations 
further support the view that a human tumor-bearing subject may to some 
extent be immunized by injections of his own tumor-tissue. 

A certain relationship has been found between the same organs in differ- 
ent species of animals, as with the testis (v. Dungern and Hirschfeld Schenk), 
and with the sexual cells of fish (Dunbar). 

Nevertheless in the sera of animals rendered immune to transplantable 
tumors there has been no satisfactory demonstration of specific cytolysins, 
agglutinins, or precipitins. Nor have complement deviation tests in such 
sera been clearly successful. 

Lambert and Hanes were able to cultivate rat sarcoma cells in the plasma 
of rats immunized against this sarcoma, although the sera of immunized 
alien animals inhibited the growth while the normal alien sera did not. 

The serological evidence therefore points to some other form of immunity 
and this is apparently to be found in the local and general hypersensitization 
produced in the immunized animals elicited by absorption of regressing or 
inoculated tumor-tissue, v. Dungern and Coca with the hare sarcoma, 
and Weil with the Flexner-Jobling rat carcinoma demonstrated anaphylactic 
reactions in the inoculated animals and a local inflammatory allergetic 
reaction to inoculation. This conception of tumor immunity is in accord 
with the fact that complete resistance to implantation of a virulent sarcoma 
may exist in animals bearing an actively growing tumor of the same sort. 




Fibroma is a term applied to tumors composed chiefly of connective 

Many of these tumors are composed exclusively of connective tissue and 
are designated as pure fibromas. More of them are covered by hypertrophic 
lining epithelium, as in papillary fibroma, or contain glandular structures, 

Fig. 25. — Benign pigmented fibro-epithelial wart. (Photo by B. H. Buxton.) 

adenofihroma. In the nerve-trunks they may show atrophic nerve-fibers 
for which, and on account of their origin, they are called neurofibroma. 
Blood-vessels may be very abundant or the lymph-vessels may be in excess, 
in angiofibroma and lymphangiofibroma. Neoplastic connective tissue is 
also observed in various other benign and some malignant tumors as in 
fibrolipoma, fibrockondroma, etc. 

The gross appearance of true fibromas is usually characteristic. They 
are circumscribed, encapsulated, usually lobulated, firm or soft tumors, and 
on section exhibit a grayish translucent appearance. The blood content 
varies but is usually slight. Secondary changes yield corresponding altera- 



tions in texture. Multiplicity is a striking feature of neurofibromas both 
in the skin {fibroma mollusciim) and in the organs. From this character- 
istic appearance there are all gradations merging toward the products of 
chronic inflammation. In some very early fibromas of breast and kidney the 
isolation of the tumor nodule is imperfect although the later stages of such 
growths are well encapsulated. There is no sharp dividing hne between 
chronic productive inflammation and fibroma, although characteristic exam- 
ples of each are widely separate. The inflammatory nodule merges insen- 
sibly with the surrounding tissue, its histology lacks the definite features 
of a neoplasm and it occurs under chnical conditions which supply the inflam- 
matory agent, as in chronic mastitis, and chronic rhinitis. Yet, as Ribbert 
remarks, the longer an inflammatory nodale persists the more in some cases 
it may approach the quaUty of a neoplasm. It must be admitted that 
chronic productive inflammation may pass insensibly into fibroma, a fact 
most frequently illustrated with some forms of local elephantiasis. 

The Structure of fibromas varies greatly but it commonly presents 
fibroblasts, fibrils, blood- and lymph-vessels. 

Soft fibromas contain a considerable proportion of cells lying in a soft 
matrix which is often edematous. Much vascularity also reduces the con- 
sistence. Hard fibromas contain fewer cells and firmer more abundant 
matrix which is often hyaline. In very dense fibromas arising from perios- 
teum and tendons very few cells may be present and wide areas of hyaline 
matrix are observed. The consistence is also affected by calcification, 
mucoid degeneration, and the formation of cysts. 

The cells of genuine fibromas are usually larger and more numerous than 
in adult connective tissue. In actively growing fibromas the chromatin 
content of the nuclei is also increased and the tissue has a distinctly neoplastic 
character. Even in some hard fibromas as in keloid, the cells may be of 
large size throughout, but in many cases the active cells may be limited to 
the periphery or to isolated foci chiefly about vessels while the main portions 
of the tumor are acellular. The arrangement of the cells is uniform and is 
determined by pressure or by the presence of originating structures as blood- 
vessels or nerve- trunks, 'or gland alveoli. 

With an increasing proportion of cells, diminution of matrix, and 
irregularity in size and shape of the cells, the tumors show more active growth 
and a tendency to recurrence, and to these tumors the term fibrosarcoma is 
given. In distinguishing fibroma from sarcoma Borst relies chiefly on irregu- 
larity in size and shape of the cells as well as overgrowth of nucleus and 
absence of definite stroma. When the bulk of cells greatly exceeds the matrix 
the tumors are usually of active growth and require careful removal to pre- 
vent recurrence. 

The matrix of fibromas varies greatly. In many cases it has the charac- 
ters of embryonal connective tissue; in keloid it is hyaUne; in dense tumors 
it may be very firm but fibrillar; and it is often edematous and infiltrated 
with leukocytes. 

The fibroglia fibrils of Mallory are regularly present and may be very long 
and quite abundant. They 'are best demonstrated in material promptly 
fixed in Zenker's fluid and stained by a mixture of anilin blue, 0.5; orange 
G, 2.0; oxalic acid, 2.0, water, 100. The bundles of wavy collagen fibrils are 
abundant chiefly in hard fibromas. Ribbert regards the matrix of fibromas 
as a product of ceU activity but not as a derivative of cell processes. The 
fibrils he considered to be the result of cell activity upon a secretory product 
which has many of the qualities of fibrin but with an altered staining reaction. 
This origin would explain the great length of many fibrils, the great bulk of 


the matrix, and the presence of "many fibrils and much matrix with very few 
cells. Elastic fibrils are usually scanty. 

Secondary changes occur in the matrix, including hyaUne transformation, 
calcification, fatty and mucoid degeneration, edema, and necrosis, and in 
connection with bone, ossification. Also the remnants of invaded structures 
may be included in fibromas, as fat-cells which may be incited to prolifera- 
tion, gland alveoli, muscle-cells and nerve fibrils. By metaplasia portions of 
fibromas may be converted into cartilage or bone. 

Blood-vessels occur in normal form and number, or as irregular thin-walled 
spaces, or as sinuses lined by endothelium alone. In angiofibroma the vessels 
are very abundant. They appear as irregular spaces and canals lying in the 
connective tissue, they are lined by a moderate number of flat or cubical 
endothelium, but the connective tissue is the predominant element. Some 
authors include among fibromas the majority of angiomas. It seems better, 
however, to separate from fibromas all very vascular tumors in which blood- 
vessels form a predominant organoid unit. Fibroma is a histioid, not an 
organoid tumor. 

Lymph-vessels are commonly present in fibromas. In keloid the lymph 
spaces are so prominent as to suggest for this tumor the term lymphangio- 
fibroma. In local elephantiasis which is closely related to fibroma the dilated 
lymph- vessels are numerous. Edema and lymph-cysts form from obstruction 
to the lymph flow. 

The natural history of fibroma is that of a slowly growing tumor progress- 
ing steadily or intermittently over many years. In many instances, as in 
neurofibroma, there is a natural tendency for the tumor to reach a certain size 
and then to remain stationary, probably from mechanical interference with 
blood-supply. In .very large fibromas as of breast, skin and ovary, hyaline, 
calcific and necrotic areas bear witness to the regressive tendency that over- 
takes this type of tumor. A spontaneous cure may thus be accomplished. 
In a large fibroma of the quadriceps tendon I have seen almost complete 
destruction and permanent arrest of growth by central softening. Many 
mammary fibromas suffer hyahne change and arrest of growth after 1 5 to 20 
years. Kosinski has reported complete spontaneous atrophy of neurofibro- 
mata. Sarcomatous transformation is very rare but occurs with neuro- 
fibroma. In the mixed adenofibroma of breast cancerous changes may de- 
velop in the epithehum. Recurrence after operation is characteristic of keloid 
but usually consists of an extension of the tumor process over formerly intact 

Origin of Fibromas. — The exact point of origin of most fibromas is still 
undetermined. That many of them arise from misplaced islands of tissue, 
according to Cohnheim's theory, is \'ery probable. Chiari observed a coii- 
genital polypoid fibroma at the perineal raphe, and Ribbert has seen a fibroli- 
poma in the same situation. Arnold described a congenital chondrofibroma 
of the skull which he referred to a disturbance in the development of the 
cranium. In several reported fibromas of the testis there were elements 
suggesting that the tumor represented a one-sided development of a teratoma. 
Certain fibromas in the adult appear in the region of embryonal clefts where 
superfluous cells may be supposed to exist. In the breast, Williams found 
evidence that 14 per cent, of the adenofibromas of this organ arise from 
embryonal rests, and the frequent association of chondroma and osteoma 
with fibroma, especiaUy in the dog's breast, favors this view. 

For other fibromas one must apply the theory of local irritation and dis- 
turbance of nutrition, as in keloid which commonly follows scarring; in 
papillary fibromas and adenofibromas of mucous membranes which are 



associated with catarrhal inflammation; and in fibromas of periosteum and 
joint areas which may follow repeated trauma. 

In a third group the clinical features point to a congenital or acquired local 
predisposition, of which multiple neurofibroma is the best example. Ele- 
phantiasis seems to exhibit a general predisposition often combined with the 
effects of recurring inflammation. 

It is especially in the second and third etiological classes that one encoun- 
ters the less definite tumor-hke processes which it is sometimes difficult to 
classify, and in which one must recognize the cumulative influence of inflam- 
mation and chronic disturbance of nutrition and the passage of inflammatory 
into self-perpetuating neoplastic processes. 

l|k„-> ,• ... vA^--- - --■-', ■ •■ Ay. I 

Fig. 26. — Fetal fibro-adenoma of breast in a girl of i8 years. The tumor grew rapidly, 
measured 9 cm., was soft, bluish, resembling sarcoma. 

Clinical Types of Fibroma. — The clinical types of fibroma are numerous 
and some of them form characteristic clinical entities which not only illus- 
trate the peculiarities of this group of neoplasms but throw light on the nature 
of tumors in general. 


Fibrom.a Molluscum. — The most frequent fibroma of the skin occurs in 
multiple form, as fibroma molluscum. Recklinghausen showed that this 
tumor arises from the cutaneous nerve filaments, a conclusion which had been 
suggested by some previous writers and has been fully verified by many latei 
studies. The peculiar clinical characters of the disease were fully described 
by KoUiker, i860; Hitchcock, 1862; Virchow, 1863; Murray, 1873; Atkinson, 
1875; Balzer, 1879; and Winiwarter, 1876; but Recklinghausen in 1882 traced 


degenerating nerve-fibers in several characteristic cases and stated that all 
these tumors arise from nerve-trunks or filaments. An extensive literature 
has accumulated on this subject and has been reviewed by Recklinghausen 
and more recently by Bruns, Sarazanes, Verocay, and Herxheimer and Roth. 

Two main types of neurofibroma are commonly distinguished, although 
both may be observed in the same patient, (i) Pigmented multiple neuro- 
fibromas, typical fibroma moUuscum, and (2) plexiform neurofibroma. 

In typical fibroma moUuscum the tumors are multiple, and they may 
become so numerous as to nearly cover the entire surface of the body. They 
may be as small as pin-head, more often of the size of a pea, occasionally they 
reach the bulk of an orange, and rarely they attain large dimensions. They 
may become pendulous (fibroma pendulum) . All portions of the integument 
may be affected from the scalp to the soles of the feet, while in some cases the 
mucous and serous membranes and internal organs are involved. In the 

skin they lie just beneath the epidermis, or beneath the derma, or they follow 
up the nerve- trunks. They are movable laterally. They are sometimes 
painful and when they are not palpable this symptom readily suggests 
erroneous diagnoses. They may affect any nerve and as a rule they are not 
symmetrically placed. Pigmentation of the skin over the tumors or else- 
where is often marked, so that many authors speak of the disease as pigmented 
neurofibromatosis . 

In structure the tumors are usually cellular and soft. The growth sur- 
rounds the nerve or involves one side or is connected with the nerve-trunk 
by a thin pedicle. The nerve-fibers show all stages of atrophy. The larger 
tumors frequently involve other structures of the skin as the sebaceous and 
sweat-glands. They may be quite vascular and many authors trace a con- 
nection between neurofibroma and many angiofibromas and nevi. They are 
sometimes associated with xanthoma (Delore, Poncet), and rarely with 
multiple lipoma (Vallas, Mouchet). Some early lesions suggest an origin 
from the connective tissue about sweat-glands, but even here the process 



has been traced to nerve filaments belonging probably to the sympathetic 

Plexiform neurofibroma (Rankenneurome. Neurofibroma cirsoides. 
Bruns). — The plexiform neurofibroma is larger and the tumors axe less 
numerous than with fibroma moUuscum. Their chief features are well 
illustrated by the case of Bruns, of a boy of i8 years who presented a pendu- 
lous fibroma 15 cm. in diameter, together with many small-pigmented tumors 
of the type of fibroma moUuscum. The grandparent and three brothers 
suffered from fibroma moUuscum. The congenital and hereditary tendency 
is pronounced. 

The tumors are composed of an agglomeration of many soft lobules 
surrounding nerve-trunks which may long be preserved and appear as 

irregularly coiled or straight nerve bands 
traversing the tumor. On dissection many 
of the coiled and thickened nerve-trunks may 
be unravelled. In neurofibroma cirsoides 
which appears chiefly in the face and scalp 
the blood-vessels are abundant. The peri- 
pheral zone is usually dense fibrous tissue, 
enclosing more ceUular connective tissue, while 
the core is occupied by the nerve-trunk. Sar- 
comatous changes have occurred in several 
cases of plexiform neurofibroma and in those 
of Genersich and Herxheimer there were 
metastatic nodules in the lung. Local recur- 
rence is very frequent. Many other pecu- 
liarities are illustrated in the cases collected 
by Recklinghausen and later writers. It is 
probable that the great majority of actively 
growing fibromas and fibrosarcomas of the 
limbs are of neural origin. 

Visceral Neurofibromatosis. — The internal 
nerve-trunks and filaments may be the seat of 
large plexiform or small multiple fibromas. In 
the intestinal form the tumors may occur at 
any point from the lips to the anus. Certain 
cases of macroglossia have been found to be 
diffuse neurofibromatosis and related to ele- 
phantiasis especially of the lymphangiectatic 
type (Virchow). Sarazanes has collected cases 
in which the tumors occurred in the tongue, 
stomach, jejunum, ileum, colon, vagina and 
bladder. In Robin's case there was a large 
tumor of the solar plexus. Lotzbeck has de- 
scribed several large tumors pressing upon the lumbar spine and arising from 
the lumbar plexus, and Czerny observed a case involving both lumbar and 
sacral plexuses. Pomorski describes a remarkable case in which multiple 
tumors projected into the pleural cavity arising from the intercostal nerves. 
Herxheimer and Roth describe a remarkable case presenting multiple 
fibromas of skin of upper half of trunk; fibromas of intercostal and lumbar 
nerves ; small sarcoma of lumbar plexus passing into a large pelvic tumor ; 
tumor-like thickenings of terminal filaments of the sympathetic and lumbar 
intervertebral ganglia; tumor mass with chromafiine cells in adrenal medulla; 
multiple subserous nodules in jejunum; miliary endotheliomas of spinal dura. 

Fig. 28. — Hard recurrent 
fibrosarcoma of popliteal space. 
Probably neurogenic. 


In a well-defined group of cases the sympathetic system is the chief or 
exclusive seat of the lesions (Czerny, Adrian). In the case of Simon and 
Heche the mesentery was much thickened and beset with many fine nodules. 

Several cases with adrenal tumors and pigmentation serve to explain the 
long-suspected relation to Addison's disease (Chauffard, Suzuki, Hoffmann, 
Kawashima, Vignolo-Lutati). The tumors of the adrenal are composed of 
medullary chromaffine cells. Gliomas of the central nervous system have 
been observed in a few cases (Verocay) . 

The exact nature of the cells involved in these tumors is not easily de- 
termined and probably not uniform. Herxheimer and Roth conclude that 
the small dermal tumors are chiefly fibromas. In larger growths they find 
that the cells of the sheath of Schwann and of the endoneurium multiply. 

^^/t *'" .'■■' '' '.l«i f^.^- "'■'"■** ■'■''iii<»' .- ,\ 

Fig. 2g. — Markedly fasciculated structure of a neurosarcoma. Neurofibroma of skin. 

Concentric groups of cells which are so characteristic of this tumor they inter- 
pret as endothelium of the nerve sheaths. In the ganglionic lesions the 
endothelium in the capsules of the ganglion-cells multiplies actively. In 
fact all the elements of the nerve filament or trunk except the ganglion-cells 
and fibrils participate in the tumor process. 

According to Verocay's conception the tissue of neurofibroma is not con- 
nective but consists of nerve-cells or their embryonal equivalents which have 
not been properly employed in the development of the nervous system. He 
therefore suggests the term neurinoma. The validity of this view depends 
on the nature of the cells of the sheath of Schwann. If these are specialized 
nerve or glia cells, Verocay's interpretation would seem acceptable. 

Elephantiasis. — It has long been recognized that there exists a close con- 
nection between multiple tumors of nerve-trunks and some forms of ele- 
phantiasis. The nature of this relation remains ill-defined but it is strongly 
suggested by the occurrence of multiple fibromas in many cases of elephan- 
tiasis, by the hereditary and congenital character of both conditions, by the 
occasional presence in both of pigmentary changes in the skin, and by the 


partial histological resemblance of the structure of the affected tissues. The 
result of recurrent inflammation and lymph stasis are the two main factors 
which separate non-parasitic elephantiasis from fibromatosis. 

Elephantiasis or diffuse fibromatosis affects the skin over large areas, as 
of an entire Hmb, or occurs in more circumscribed form when its neoplastic 
characters are more distinct. It is acquired and sporadic or congenital arid 
endemic. Beginning early it produces a soft mucinous thickening Or in 
later periods the tfssue is firmer or sclerosed (elephantiasis moUe and durum). 
The lymph-vessels are commonly increased, enlarged, or may be cystic (ele- 
phantiasis lymphangiectatica) . 

Hygroma cysticum (Wernher) is a circumscribed form of congenital 
lymphangiectatic elephantiasis appearing in the neck, sacral and perineal 
regions of fetuses, infants, and children. The blood-vessels are sometimes 

" ■'=• ^' ''^ ^''- 

Fig. 30. — Neurofibromatosis. Detail of process in nerve-filaments. In the center 
proliferation of sheath-cells; in intermediate zone, endothelial cells; in peripheral area, 
fibroblasts. {After Herxkeimer.) 

much increased (elephantiasis telangiectatica, elephantiasis angiomatosa). 
In some remarkable cases the skin of the entire limb was greatly swollep, 
dark blue, and composed of many thickened varicose veins often containing 
phleboliths (Rokitansky, Pitha, Leisrink, Unna). All structures of the skin 
may be involved and there is increase of connective tissue diffusely and about 
nerves, vessels, and glands. The sweat-glands may be hypertrophic. The 
epidermis may be thickened, pigmented and corrugated, and in the remark- 
able form of congenital ichthyosis it is hypertrophic and extensively horni- 
fied. In cases reported by Naegeli and by Morgan elephantiasis was associ- 
ated with tumor-like thickening of the nerve-trunks supplying the area. In 
several cases reported by P. Bruns the elephantiasis was associated with 
definite neurofibromatosis. In certain cases usually congenital the thick- 
ened skin falls in overlapping folds like drapery (Mott) and for these cases 
the terms pachydermia or pachydermatocele are employed. Jordan dis- 
tinguished two types of congenital elephantiasis, one of which is chiefly a 


diffuse neurofibromatosis, the other involving nerves, other cutaneous struc- 
tures, and the muscle-tissue. In each case the blood-vessels gave origin to 
the new tissue. 

In most cases of acquired (elephantiasis the nerve-trunks play a subordi- 
nate part and the process involves all the cutaneous structures, yet in many 
cases of long-standing neurofibromatosis similar extension of the hyperplastic 
process is observed. The clinical history indicates that this ' extension is 
chiefly owing to recurrent inflammation and lymph stasis. Thus the disease 
commonly progresses with repeated attacks of erythema or erysipeloid, and 
it may develop in the course of chronic suppuration. These observations 
early led to the assumption that elephantiasis was of infective origin and 
confusion arose between elephantiasis arabum (diffuse fibromatosis) and 
elephantiasis graecorum, which was true leprosy. Carter has pointed out 
many points of resemblance between the two types, especially in the nerve 
lesions, and shown that true elephantiasis may compHcate leprosy. To-day, 
both leprosy and filariasis may be readily separated from the other causes 
of elephantiasis. 

These chnical relations of elephantiasis as well as the histological structure 
demonstrate that the process is not a true neoplasm but belongs in the class 
of fibromatoid reactions of predisposed subjects to chronic disturbance of 

Neuropathic papilloma (Gerhardt), Nevus unius lateris (Barensprung), 
Nevus linearis (Unna), Nerve nevus (Simon). — This lesion occurs as multi- 
ple flat or bean-shaped, warty outgrowths of hypertrophic dermal papillae. 
They follow the distribution of definite cutaneous nerves like herpes zoster 
and may be more numerous along Voight's lines bounding the distribution 
of the nerves (Philipson) . They may occur on any portion of the body and 
even in the buccal mucosa (Pott). In some cases they are extensively pig- 
mented (Schonberg). 

The condition is hereditary (Haegele), and often congenital, but may ap- 
pear as late as the 27th year. The onset may be rather acute as in Beigel's 
case, but the course is chronic, reaching a standstill or spontaneously re- 
gressing (Neumann). Pain may be pronounced, and paresis and paralysis 
may develop. 

The histology shows an acute stage followed by chronic connective over- 
growth of dermal papillae, but the structure varies greatly and there are no 
reports to indicate that the process is a true tumor. In Elliot's case there 
was cystic distention of the cutaneous glands. The overlying epithehum 
may be hypertrophic. Recklinghausen states that the growth affects the 
terminal filaments of the nerves, that it follows neuritis and is related to 
ichthyosis with which it may be associated. Unna doubts this relation. Al- 
though the condition is vaguely defined it appears to represent most clearly 
the influence of nervous disorder upon fibromatosis. 

General Etiology. — In the origin of the foregoing types of cutaneous 
fibromas, many complex factors seem to be concerned, most of which center 
in the nervous system. Together they constitute a somewhat definite clin- 
ical entity or dyscrasia to which the term neurofibromatosis may be applied. 
The congenital and slightly hereditary character is prominent in all the forms 
of fibromatosis, suggesting the existence of abnormal qualities in the connect- 
ive tissues of these subjects. The unfolding of the dyscrasia often proceeds 
in a somewhat orderly manner. Thus Debove observed pigmented placques 
at birth, small tumors of the skin at five years of age, and tumors of the nerves 
at eight. Yet in most cases the pigmentation and the tumors appears simul- 
taneously at a later age, 10 to 30 years (Marie). French authors, especially 


Sarazanes, emphasize the nervous symptoms observed in many cases, and 
point to the pain, anesthesia, defects of special senses, neuroses, epilepsy, 
mental disturbances, idiocy, fragility of the bones, and infantihsm and mal- 
formations, as evidence of general and especially of nervous deficiency in the 
subjects of neurofibromatosis. Various digestive disturbances, jaundice, 
urobihnuria, and alterations of nitrogenous metabolism have been cited in 
support of the theory of general dyscrasia. Feindel and Oppenheim describe 
marked forms of neurofibromatosis, in which the general symptoms are 
present with punctate or lenticular pigmented spots in the skin while 
the tumors are absent, scanty, or appear later. 

For general fibromatosis various theories of origin have been assumed, as 
the theory of recurrent or chronic infection as suggested by the recurrent 
inflammatory processes in elephantiasis and its resemblance to leprosy. The 
theory of auto-intoxication is suggested by many constitutional symptoms 
and especially by certain cases of neurofibroma which clinically resemble 
Addison's disease (Chauffard, Bernard). Feindel points out that the pig- 
mentation and tumors may date from pregnancy, and alcoholism and over- 
use of arsenic have been reported as exciting causes. Brickner has de- 
scribed in pregnancy a peculiar punctate pigmentation of the skin which 
seems to fall with neurofibromatosis. It is generally assumed that the vic- 
tims of fibromatosis are the subjects of a congenital malformation of the ec- 
toderm which under a great variety of exciting causes may slowly or rapidly 
develop one or more of the manifestations of the disease. The great mass of 
data that has accumulated in this field seems to call for a general assent to 
this hypothesis, but there is danger that the conception of the scope of this 
dyscrasia may be extended over too many pathological phenomena. 

Keloid. — Keloid (kijXjj, tumor) or Cheloid (xvM, claw) is a peculiar 
overgrowth of hyaline connective tissue developing in the skin of predisposed 
subjects after trauma or scarring. The traumatic origin is usually obvious, 
as in the extensive scarring of burns. More localized keloids occur in which 
it is difficult to trace the probable traumatic origin and these are called 

The subjects of keloid may exhibit a special predisposition to overgrowth 
of connective tissue after injury and in the negro this predisposition is racial. 
The character of the trauma is diverse. Extensive burns by heat or escha- 
rotics are specially prone to develop keloid, but needle punctures (Gumbel) 
scratches, vaccination, repeated pressure of heavy weights and surgical 
wounds, have preceded the condition. 

The typical keloid has ,a characteristic structure and consists of anasto- 
mosing strands of thick, hyaline, collagenous matrix between which lie many 
well-nourished fibroblasts. Small well-developed blood-vessels are present, 
and lymph spaces may be so prominent as to lead to the use of the term 
lymphangiofibroma for some cases of keloid. Elastic fibers are missing. 
Glandular structures of the skin are pushed aside and this fact leads Goldman 
to conclude that the growth of the tumor is chiefly central. The periphery is 
not well defined but is often more cellular with thinner fibers than the cen- 
tral pottions. Yet this feature is not so pronounced as tordeprive the growth 
of its circumscribed character or to invalidate the statement of Ribbert 
that keloid grows largely from its own resources. Secondary changes are 
chiefly those of increasing hyalinosis. 

Keloid occurs in the form of a broad flat elevated firm thickening of scar 
tissue which gradually extends over a considerable area, or as rnore circum- 
scribed projecting or even polypoid outgrowths. The overlyiljg epidermis 
is usually, shiny, atrophic and separated from the tumor by ajl;iiin layer of 


cutis. The natural history is of a slowly progressive growth which tends to 
inyolve the whole area of affected scar tissue, or remains hmited to the seat 
•of^a trauma. After extirpation it often develops anew in predisposed tissues, 
but never produces metastases. It may be made to disappear under 

The neoplastic properties in keloid are not pronounced. Especially at 
Its inception it belongs in the group of fibrous overgrowths which depend upon 
chrome disturbance of nutrition. Yet it sometimes possesses considerable 
momentum of growth, its cehs exhibit nuclear hypertrophy, and especially 
in the so-called spontaneous circumscribed forms, it may sometimes be re- 
garded as an imperfect neoplasm. The remarkable case of Porter's in which 
a very bulky keloid of face and arm developed in ten months after a burn 
suggests both the neoplastic nature of the process and a predisposition to 
the disease. 

Fig. 31. — Tumor-like keloid following needle puncture. Remained stationary after a few- 
months of rapid growth. (After Breniz'.r, A. S. 61.) 

Xanthoma. — Xanthoma, or xanthelasma as named by Wilson and Smith, 
is a common brownish-yellow tumor of the skin appearing most frequently 
in the eyelids (92 percent., Poensgen) but occurring in multiple form in other 
regions (xanthoma multiplex, generalized xanthoma). It may appear 
first in the eyelid, later in other localities (Korach) and like multiple 
fibroma often follows the course of cutaneous nerves, especially along Voight's 
fines (Kobner), and symmetrically along folds of the skin. Apparently 
similar growths have been described in the respiratory and pharyngeal mu- 
cosa, the serous membranes, endocardium, tendon sheaths, on the cornea, 
and in the pancreas in diabetes. According to Unna generalized xanthoma 
is sharply separated anatomically from xanthoma palpebrarum. 

In the skin the growth is flat (xanthoma planum) or beanshaped (xan- 
thoma tuberosum) or polypoid and pendulous. Usually of small size, they 
occasionally reach considerable dimensions (Poensgen, Ehrmann). They 
are often congenital, but rather more frequent in adults and old persons. 
It was shown by Chambard that general xanthoma may develop in the course 
of icterus or with other diseases of the liver and with diabetes (xanthoma dia- 
beticorum). Not a few subjects of early xanthoma later develop diabetes, 
yet Johnston and Torok separate the skin affection in diabetes from the com- 
mon xanthoma vulgare. The diabetic lesion may be transitory while the 
other is permanent. 

In sorfle'cases xanthoma tuberosum is associated with nevi, vascular and 


pigmented, illustrating a congenital predisposition, and this fact may support 
Borst's conclusion that the two conditions are anatomically related. Xan- 
thoma may also be associated with multiple lipomas (Ehrmann). Vir- 
chow regarded xanthoma tissue as intermediate between connective and fat 
tissue and employed the term fibroma, believing that the cells 
were derived from fibroblasts. The histological and clinical features point 
to the existence of a peculiar formation of lipoids in the cells of a fibroma under 
the influence of a local nervous and a general metabolic disorder. In 
xanthoma palpebrarum the deposit of lipoids precedes the tumor growth 
and first appears as yellow streaks along the blood- and lymph- vessels, and 

The structure of xanthoma is characterized chiefly by the great abundance 
of lipoid globules in large swollen polyhedral and smaller spindle-shaped cells. 
This material is an orange yellow lipochrome; more of it is stained by Sudan 
III than by osmic acid, and Stoerk, with Pinkus and Pick, have showed that 
the globules are doubly refractive and of myelin nature or probably a choles- 
terin fatty acid ester. The central cells may become greatly distended 
with this material, while the peripheral cells may show the spindle shape 
of growing fibroblasts. Much fatty detritus may accumulate in these tumors 
and giant-cells similar to those in epulis and containing or surrounding fatty 
crystals or masses may form (Touton) . In such cases the tumor may resem- 
ble a spindle- and giant-cell sacroma. I have seen such a tumor from the 
peroneal tendon sheath of a diabetic subject who also showed xanthoma 

Waldeyer early showed that the first changes in xanthoma palpebrarum 
consist in the appearance of fatty globules in certain perivascular groups of 
star-shaped fibroblasts. Many authors, including Borst, trace the xanthoma 
sells to the lymph endothelium and he also finds a close structural resem- 
blance between some xanthomas and pigmented nevi. Hallopeau and Torok 
consider the cells as equivalent to embryonal fat-cells which they may 
strongly resemble. Waldeyer found many plasma-cells about the vessels 
of the eyelid and believes that these become infiltrated with fat. Hebra, 
Wilson, Geyer, and others have observed a proliferation of the sebaceous 
glands, and Knauss of the hair follicles, sweat-glands and blood-vessels. 
Pollitzer in cases of xanthoma palpebrarum traces the fatty cells to degen- 
erating fibers of the orbicularis muscle, and he would separate this lesion 
from xanthoma multiplex, which he regards as a fibroma or keloid undergoing 
fatty degeneration. Unna also traces a close connection between the 
fatty cells of early xanthoma palpebrarum and the muscle- fibers of the orbicu- 
laris which lies close to the skin, but he considers the fat to be a deposit 
from the lymph- vessels. 

It thus appears that many different anatomical elements may be concerned 
in the clinical lesions of xanthoma, and that not all of the clinical forms 
have a relation to tumors. Some of them, however, are certainly neoplastic 
and fibromatous. Since the fatty change seems to be identical in nature 
wherever it occurs there would seem to be room for the term xanthomatosis 
as applying to this process wherever it occurs and without limitation to a 
specific anatomical lesion. 


Tumors of this class are comparatively rare, but of very general distri- 
bution. From the maxillse, especially the superior, fibromas arise and reach 
moderate dimensions, projecting into the mouth with ulceration of the mu- 



cosa, or beneath the skin. In the interior of the nlaxillae fibroihas may develop 
from disturbances in the formation of the teeth (Blauel). The long bones are 
occasionally the seat of hard fibroma for which a history of trauma is common. 
[n the fasciae especially about the joints are encountered hard fibromas which 
may reach a considerable size. The osteal and periosteal fibromas frequently 
exhibit areas of calcification, chondrification and true bone formation, in 
which processes characteristic osteoblasts engage. Bone also forms in fi- 
bromas apart from any connection with periosteum. In the large fascial 
fibromas hyalinosis, calcification and central softening may occur leading to 
arrest of growth. 


Breast. — The fibromas of the breast are usually complex fibroadenomas 
and will be considered under epithelial tumors. 

Kidney. — Small fibromas of the kidney are of frequent occurrence. 
They are single or multiple, often bilateral, hard, opaque, homogeneous 

Fig. 32. — Fibroma of ovary. 

nodules which project under the capsule, or appear in the cortex or more 
frequently in the medulla at the bases of the pyramids. Many of them con- 
tain smooth muscle-fibers and in the cortex they usually inclose compressed 
or cystic renal tubules. These small tumors arise from disturbances of de- 
velopment of the kidney. According to Genewein they are not true neo- 
plasms but tumor-hke nodules resulting from superfluous tissue (Albrecht's 
hamartoma) . 

Large fibromas occasionally grow from the capsule or in the hilus of the 
kidney, but they are generally associated with fat and smooth muscle-tissue, 
and exhibit myxomatous changes. Under the term myxoma fibromatosum 
TiUmann described a large tumor weighing 10 lb. which he extirpated from a 
woman of 28 years and which soon recurred. Claus, Wilks, and Busse and 
Reinach have reported large tumors of capsule, cortex, or hilus of the kidney 
which were nearly pure fibroma or mixed with lipoma, myxoma, or myoina. 
In a large tumor of this class I found lobules of nearly pure fibroma, with 


others of fat and myxomatous tissue. These growths must owe their origin to 
some developmental disturbances of the kidney, the nature of which is un- 
determined. Small fibromyomas or nearly pure multiple leiomyomas are 
not infrequently seen and have been described by Lartigau and Larkin. 

Testis. — Pure fibroma of the testis is very rare (Langhans), but I find 
that tumors chiefly fibromatous have been described as occurring in the tunica 
albuginea, testis, epididymis, or cord. On close examination most of these 
tumors disclose evidence of complex composition, as in Chevassu's case. 
They have sometimes reached a large size. Their origin has been referred to 
the interstitial tissue of the affected locaUty, but that some of them are re- 
lated to teratoma is very probable. 

Fibroma in the ovary presents several features of interest, {a) The dif- 
fuse tumors arise chiefly at the lateral pole and involve much or all of the or- 
gan. They are capable of reaching enormous dimensions. A frequent 
accident is strangulation by torsion of the pedicle. The structure varies 
greatly in richness in cells, or vessels; from admixture with myoma; from sec- 
ondary changes toward myxoma, osteoma, chondroma; and from necrosis, 
and cystic softening. Adfer describes an ovarian intracanahcular adeno- 
fibroma. Ovarian hematoma may be connected with their origin (Brothers), 
and a special type develops about corpora lutea (Rokitansky). Others may 
be of teratomatous nature. 

(&) Papillary fibroma appears as multiple warty outgrowths of one or 
both organs. 

(c) Ovarium gyratum is a form of superficial fibrosis of both ovaries which 
obhterates the cortical structures and transforms the organ into an enlarged, 
hard, very opaque mass, with wart-hke surface. The condition was so named 
by Adler and is not infrequent. 


Although there has been much debate over the relation of nasal polyps 
to true neoplasms, here as elsewhere it is impossible to draw a sharp hne be- 
tween the products of chronic inflammation and tumor processes, so that 
there is some justification in the current usage by rhinologists of the term 
fibroma for these common growths (Bosworth, Lit.). Nevertheless, it is 
quite clear that in the nares more than in any other mucus membrane the 
polypoid outgrowths of chronic inflammation lack the histological features 
of an autonomous new growth. In fact, as Chiari claimed in 1887, many of 
them consist of nothing more than localized edematous areas of mucous 
membrane rendered protuberant by mechanical means, but without other 
changes. Once estabhshed, however, these masses are subject to various 
grades of hyperplasia of their elements which render them not only persist- 
ent, but often progressive, and in such cases there may be considerable 
change in the appearance and proportions of various cells. Since this change 
is seldom pronounced, the groups of nasal polyps must stand among the 
purest examples of pseudoneoplasms of inflammatory origin. 

Nasal polyps are probably always preceded by chronic rhinitis and 
Tissier traces an unbroken series of cases from simple chronic rhinitis through 
hyperplastic rhinitis to polypoid inflammatory outgrowths. The tumors 
appear chiefly in young subjects and infants, rarely after 30 years, generally 
at the ostia of the mucous sinuses opening into the nares. Empyema of 
these sinuses is a common antecedent. They are commonly multiple, 
sometimes very numerous, and they occasionally reach a large size, pressing 
on adjacent structures, and widening the bridge of the nose, with obstruction 


of passages and various anatomical and clinical sequelae. Heymann pictures 
a very large nasal polyp which filled the nasopharynx. They are soft in 
the early stages, firmer if old, gorged with blood, or gray and translucent 
from edema, with smooth warty or papillary surface. They may begin 
as numerous low elevations of the mucosa, but soon tend to become 

Histologically three somewhat different varieties are described by Zarniko, 
fibroma edematodes simplex, adenofibroma edematodes, and fibroma edematodes 

Fibroma Edematodes Simplex. — This tumor is composed of loose con- 
nective tissue with few or many spindle- or star-shaped fibroblasts. The 
stroma contains loose collagen and some elastic fibers. The grade of edema 
varies but is usually pronounced, sometimes leaving wide spaces, or if 
slight, as in old cases, yielding rather firm cellular areas. Lymphocytes 
may be abundant, beneath the epithelium, about the vessels or in small foci. 
Eosinophile cells, Charcot-Leyden crystals, and even extensive pigmentation 
may form the sequel of hemorrhages. Nerve-fibers losing themselves in 
the stroma have been demonstrated by Billroth and by Kalischer. Blood- 
vessels are abundant especially in the pedicle but tend to disappear. In 
some old cases they are very numerous and cellular suggesting angioma. 
Dilated venous sinuses add to the bulk of many cases. Lymph sinuses are 
less numerous. Areas of calcification and islands of true bone suggest an 
influence of the neighboring periosteum. The lining epithelium is hyper- 
trophic, ciliated, or transitional, or squamous from attrition. The edematous 
fluid is an albuminous transudate, but may contain considerable traces of 
mucin. Hyahne globules may also be present. 

Adenofibroma Edematodes. — In some polyps new gland formation is 
prominent and much of the tissue is composed of various slightly distended 
alveoli lined by several layers of hypertrophic epithelium. Goblet cells 
may be very numerous. In some cases the epithelial hyperplasia is pro- 
nounced but it rarely suggets a true neoplasm. 

Fibroma Edematodes Cysticnm. — Most of the large and some smaller 
polyps owe their bulk to the formation of cysts which arise from distended 
alveoh lined with overgrowing cells, or if ducts are occluded, with flat cells. 
The contents are serous fluid, mucus, pus, degenerating epithehum, fatty 
detritus, or hyaline or calcific deposits. 

The derivation of aU these features and many other details in the structure 
of nasal polyps have been fully traced by Heymann, Zuckerkandl, Wright, 
Zarniko and many others. 

The natural history of nasal polyps is that of a persistent and slowly 
progressive pseudoneoplasm. They seldom spontaneously disappear, except 
through accidental expulsion. They are frequently found as a complication 
of malignant tumors, but malignant transformation of a benign polyp is 
extremely rare, Heymann finding only three somewhat doubtful cases. 
They do not recur after complete removal which is sometimes difficult, but 
new tumors often succeed those removed. In the etiology a special pre- 
disposition may possibly exist, but the growths are adequately explained as the 
result of chronic inflammation and the irritation of retained secretion. 

Choanal Pol}^). — Arising from the posterior nasal orifice very cellular 
firm tumors occur which lack the edema of the intranasal growths, but yet 
attain large dimensions, reaching into the pharynx or, as in Storck's case, 
to the larynx. Zaufal removed such a tumor weighing 112 gm. Zarniko 
observed extensive bone formation in one case with marked hyperplasia 
and metaplasia of the epithehum. Some of these growths show more definite 



neoplastic characters than the intranasal growths. The choanal polyp 
often originates from the mucosa of the accessory sinuses, maxillary and 
antral, ethmoidal, or sphenoidal. These tumors are characterized by a 
tendency to reach a large size, by the presence of a long pedicle, and by recur- 
rence after removal of the tumor without the pedicle. After repeated 
recurrence it is usually found necessary to open the sinus and attack the 
growth at the base, and many rhinologists employ the radical method at 
once in dealing with these polyps (Moore, Syme) . 

Juvenile Nasopharyngeal Fibroma. — This rare and pecuhar tumor occurs 
chiefly in males and almost invariably between the ages of lo and 25 years. 
It is a very firm almost cartilaginous tumor which appears in the vault 
of the pharynx and grows in several definite directions. According to 
Bensch it produces an intrapharyngeal tumor when arising from the basilar 




■'^ 'f 

' ., j' art 

Fig. 33. — Malignant angiomyxoma growing from superior nares in a child. 

fibrocartilages, the upper cervical vertebrae, or the internal lamina of the 
pterygoid process; or an extrapharyngeal growth when arising from the 
cartilage of the foramen lac. ant., or the sphenopalatine fossa. The intra- 
pharyngeal tumor extends forward into the nares and the adjacent sinuses, 
causing atrophy of the bony structures. From the sphenoidal origin the 
growth extends down between the masseter muscle and the mucosa; or it 
pushes between the pterygoid and styloid muscles into the temporal fossa 
and forward into the malar region; or through the inferior orbital fissure 
it extends into the orbit, or by way of the superior orbital fissure or lamina 
cribrosa it reaches the cranial cavity. From these points the courses from 
the two seats of origin overlap. 

In structure the tumor is composed of dense fibrous and elastic tissue 


which rarely shows calcification, cartilage or bone. The cells are round, 
spindle- or star-shaped fibroblasts which are rather scanty except in certain 
foci of young connective tissue where they may be so numerous as to suggest 
fibrosarcoma. Mast cells and plasma-cells may be present. The vessels 
are numerous (pharyngeal angiofibroma) and sometimes cavernous. Involu- 
tion changes follow thickening and hyalinosis of vessel-walls, the stroma 
becoming hyaline and the tumor undergoing necrosis or fatty degeneration 

The course is of an actively growing tumor which disturbs various 
functions by pressure, leads to anemia from hemorrhage, suffers ulceration 
and local infection, and may prove fatal in this way, or from cerebral dis- 
turbance. Metastases are not observed. A remarkable feature of this 
tumor attested by many observers is its complete spontaneous regression 
after partial removal (Bensch, Griinwald, Konig, Bruns, Zarniko). This 
event seems to occur chiefly toward the end of the period when the tumor 
may develop, i.e., the asth year. There is thus illustrated a form of natural 
immunity, which may be referred to natural anatomical changes at the point 
of origin from which alone the nutrition of the growth is maintained. Ac- 
cording to Bensch the development of the male and female face and skull 
at pubert;^ explains the predominance of the tumor in males, while the 
completion of the cranial development at the 25 th year determines the 
spontaneous disappearance of the tumor. It is not clear that this tumor 
is always of fibromatous structure. Naab observes that chondroma and 
sarcoma may appear under much the same conditions. I have seen a tumor 
corresponding in many particulars to the above description but showing 
the structure of a sarcoma with indistinct chondromatous qualities. Con- 
fusion with chondroma seems possible. Many of the tumors are myxo- 
matous or myxosarcomatous and these recur persistently and are commonly 


Myxoma is a tumor composed of mucous tissue. 

Primary myxoma, a tumor which probably arises from embryonal mucous 
tissue, is rare, but mesoblastic tumors such as fibroma, lipoma, and chondroma 
which contain myxomatous portions are not infrequent. It is often difficult 
to determine whether the myxomatous portion of a complex tumor is origi- 
nally of this type or represents a degeneration of the more adult tissue. In 
the former case one employs the terms myxofibroma or myxolipoma, in 
the latter case the myomatous degeneration is indicated by the suffix myxo- 
matodes as chondroma myxomatodes. 

When mucous tissue is present in a neoplasm three possibilities as to its 
significance must be considered. The mucous tissue may be developed 
directly from embryonal mucous tissue and the tumor may be a prirnary 
myxoma. Or the tissue may represent a metaplastic product of other tissues 
and the tumor is a secondary myxoma. Or the tissue may result from chronic 
edema of other structures and represent a form of spurious mucous tissue, 
such as is seen in edematous nasal polyps and in atrophic fat tissue. These 
theoretical deductions cannot always be established in the practical examina- 
tion of tumors. 

Since mucous tissue does not exist in the adult body a primary myxoma, 
according to the above definition, must be rare and always embryonal. 
Fully developed mucous tissue exists in the embryo only in the umbilical 
cord and although a few myxomas of the navel have been described which 
probably arose from cord tissue the chief source of myxomas must be sought 
elsewhere. Such a source may be found in the early undifferentiated con- 
nective tissue of the embryo which has a mucous quality. Such mucous 
tissue is widely distributed in the embryo especially in the subcutaneous 
areas. It is genetically related to fat tissue into which it is extensively 
transformed during normal growth. It may be assumed that many primary 
myxomas arise from islands of such embryonal tissue. Since such islands of 
mucous tissue may readily be associated with cartilage and fibrous and fat 
tissues, which are normally developed from the embryonal mesoblast, it is 
reasonable to explain the occurrence of many myxohpomas, fibromas and 
chondromas by assuming their origin from islands of tissue which are partly 
differentiated, or which become so during the progress of the growth. It 
does not appear that pure myxomas ever tend to differentiate into fibroma 
or lipoma. It is not so easy to decide to what extent other mesoblastic 
tumors become transformed into true myxoma. Fibroma, lipoma, and 
chondroma undoubtedly become transformed by degeneration or imperfect 
growth into tumors containing areas of typical mucous tissue. The group 
of secondary myxoma develops in this way. 

Virchow considered myxoma to be very closely related to lipoma and as 
occasionally arising from embryonal fat tissue. Certain myxomas of bone 
he believed might arise from cartilage. Ribbert, however, regards such sec- 
ondary changes as spurious myxomas, on the ground that already differen- 
tiated tissues cannot become transformed into true mucous tissue. This 



objection seems to me adequate to separate the primary from secondary 
myxomas, but not to eliminate the latter class from the general category of 

In the sense that primary myxomas are foreign to the type of tissue in 
which they arise they may be regarded as always heterologous. Yet this 
character is not equivalent to that of many other heterologous tumors which 
arise from tissues widely displaced and genetically unrelated to the tissue in 
which the tumor occurs. 

Myxoma is to be sharply distinguished from epithelial tumors undergoing 
mucous degeneration. Here the mucus is a form of hypersecretion, while in 
myxoma it is an integral part of the living tissue. 

Anatomical Characters. — Myxomas are soft, lobulated, polypoid, or 
papillary tumors, which on section are smooth, translucent and gelatinous. 
The lobulation is determined by the growth about blood-vessels, which v. 
Rindfleisch finds on injection are of relatively large caliber with few capil- 
laries. _ The encapsulation of myxomas is seldom complete, the mucinous 
material tending to infiltrate surrounding tissues. On this account complete 
extirpation is sometimes difficult and recurrence follows from persistence of 
widely disseminated mucous material and cells. In some cases the condi- 
tions suggest a transformation of normal tissues into myxoma under the in- 
fluence of the mucinous material. Virchow observed the gradual extension 
of the myxomatous tendency from a recurrent tumor of the ulnar nerve over 
the greater part of the brachial plexus. 

On extraction with weak alkali myxomas yield a solution of mucus which 
is precipitable by acetic acid. Alcohol precipitates mucus in membranous 
or reticulated form resembling fibrin. 

The structure of myxoma presents typical features of spindle- and star- 
shaped cells with processes anastomosing or disappearing in the matrix, 
and cytoplasm usually containing fatty and watery droplets. Hydropic 
degeneration may be extreme in edematous tumors. 

The cells may be grouped chiefly about blood-vessels which when abun- 
dant may call for the designation "telangiectatic," or, "cavernous." Or 
they may be more numerous in the capsule in cases showing peripheral 
growth. Various admixtures of connective tissue, fat, cartilage, or bone 
occur and may deserve suitable recognition in the terminology. Cystic 
myxoma results from local edema, or hemorrhage, and in one definite group 
of cases dilated lymph spaces or vessels yield cysts. Very cellular myxomas, 
pure or combined with other tissues, and primary or secondary, constitute 
the important group of myxosarcomas. 

The matrix of myxomas varies greatly in bulk and in structure. It may 
be very abundant widely separating the cells, and appearing as nearly homo- 
geneous or faintly fibrillated or granular but always basophilic material. 
In rapidly growing tumors and in cellular areas it may be very scanty. The 
bulk of the matrix depends much on the capacity of the mucinous material 
to absorb water. In the secondary myxomas the gradual transformation of 
fibrous tissue, fat, cartilage and bone may often be traced, the matrix of these 
tissues becoming transformed into mucous tissue and the cells assuming 
spindle and star shapes. Similar changes may sometimes be observed on 
the edges of infiltrating myxomas but the invaded tissue is usually passively 
displaced by the tumor. 

EpitheHal and endothelial cells may become incorporated in the advancing 
myxoma and rendered indistinguishable from the tumor-cells. In some endo- 
theliomas myxomatous changes may be so complete as to greatly obscure 
the true nature of the growth. The myxosarcomas exhibit many more 



cells, usually more vessels, and less matrix than the simple myxomas. As a 
rule the mucous tissue is associated with areas of fine spindle-cells often 
grouped about blood-vessels. In a bulky myxoma of the fascia of the thigh 
I found that all the spindle-cells could be traced to the walls of very numerous 
capillaries. This tumor was strictly a myxo-endothehoma. Myxomatous 
tissue is often combined with fat tissue, cartilage, bone, or other neoplastic 
elements. When the formation of mucous tissue in such a tumor is the pre- 
dominating feature the growth may be designated as myxosarcoma, but 
when mucous degeneration affects only small areas the term sarcoma myx- 
omatodes is employed. The former group constitutes a numerous class of 


Pig. 34. — A vascular myxoma of superior nares. 

tumors among which the lipomatous and chondromatous varieties are most 
numerous. The malignancy of myxosarcoma depends chiefly on its local 
effects, metastases not being very common. According to Malherbe, myxoma 
of striated muscle, fasciae, and of certain viscera is usually malignant and 
equivalent to sarcoma, while in the nares, breast, bone, and nervous system it 
is almost always benign. 

The clinical course of myxoma is of a slowly growing tumor which pro- 
duces no symptoms except local swelling and pressure. After complete 
extirpation they do not recur, but thorough removal is not always readily 
accomplished and local recurrence of this benign tumor is not uncommon. 
In the skin and nerve-trunks they are frequently multiple and new tumors 
develop after extirpation of the old. Yet Virchow refers to two cases of 
myxoma of cheek, and of labium, which recurred after operation and eventually 
produced many myxomatous metastases. As a rule, malignant myxomas 
fall readily in the class of myxosarcoma. 

Simple myxoma rarely attains a large size and the onset of myxomatous 



changes usually marks a partial limitation of the growth capacities of a tumor. 
Yet the myxolipomas grow to very large size and for some of the diffuse forms 
of this tumor Virchow suggested a relation to, elephantiasis. Such a case 
affecting the leg and thigh is reported by Barling. In certain situations the 
pressure symptoms of myxomas become serious (nares, brain, cord). 

-The location of myxomas is chieiiy the subcutaneous and intermuscular 
tissue of locaUties where the structure is loose and much fat is commonly 
present. The regions most affected are the thigh, neck, cheek, leg, and peri- 
toneum, and less frequently, the bones, meninges, nerve-trunks, mucous and 
serous membranes, and hilus of the kidney. I have seen a myxoma as large 
as an orange involving three lobes of the placenta. 

Fig. 35. — Myxoma of portion of placenta. 

Certain clinical forms of myxoma are rather well defined. Myxoma of 
the nervous system affects the peripheral nerves, meninges and brain tissue. 
In the nerve-trunks single and multiple primary myxoma occurs under much 
the same conditions as multiple fibroma and probably represents a more 
embryonal form of this tumor. Some of the cutaneous myxomas are prob- 
ably of neural origin. They arise in the endoneurium and perineurium 
probably from foci of embryonal tissue, and produce rounded or elongated, 
soft but often painful tumors. After extirpation new tumors may develop. 
Wilms observed a large cystic myxoma of the ulnar nerve. 

The optic nerve is especially susceptible to myxoma, which here tends to 
pursue a characteristic clinical course (Sattler, Parsons, Salzmann). The 
tumor occurs chiefly in young subjects, produces a fusiform swelling of the 
whole nerve-trunk, blindness results, but the eyeball is usually not invaded. 



and extirpation of the tumor is usually successful. In a case studied in this 
laboratory with Knapp the central portions of the tumor were softened and 
cystic, the nerve-fibers were nearly all destroyed, while the capsule presented 
the structure of an angiosarcoma. In the meninges several soft myxomas 
have been described, especially by the older authors, of which Virchow col- 
lected a series. In the brain tissue Virchow described as myxoma certain 
tumors which were probably gelatinous gliomas. 

In the mucous membranes polypoid outgrowths in many regions present 
some of the features of myxoma. They are chiefly edematous fibroadenomas 
or inflammatory hyperplasias and contain little or no mucin. 

In the bones Virchow described both primary and secondary myxoma. 
The primary tumor he derived from the mucous tissue of the bone-marrow 
and believed that from this origin the tumor caused absorption of the bone 
and distended or ruptured the periosteum. A variety of secondary m} xomas 
arise in the bones and represent mucous degeneration of fibroma, chondroma. 

Fig, 36.- 

-Lobulated myKoma o£ right auricle, 
tumor. D, elongated portion. V, 

(After Ribbert.) A and B, Main lobes of 
vena cava sup. T, tricuspid. 

and osteoma. They may be surrounded by a bony shell laid down by 
the periosteum and portions of the original bone tissues may be found in the 
substance of the tumor. 

Congenital myxoma or myxosarcoma of the navel is described by 
Kaufmann with the report of a case and review of the literature. I have 
observed a cystic myxoma of the umbilical cord and navel measuring 10 cm. 
in diameter, and associated with congenital malformation of the kidneys. 

Myxoma of the Heart. — The endocardium is the seat of a series of tumors 
which include small fibromas, larger soft myxomas, and more cellular 
myxosarcomas. The interpretation of any one of these tumors should be 
influenced by our knowledge of the entire group. 

Brenner collected 33 reports of cardiac myxomas, of which 20 were located 
in the left auricle, chiefly on the septum, 10 on the heart valves, 2 in the 


right ventricle, and one on the apex epicardium. The smaller tumors offer 
little difficulty of recognition since they usually take the form of typical 
fibromyxoma. With increasing growth they assume a lobate, papillary, 
polypoid or even villous form and the structure is greatly altered by edema, 
hemorrhage and progressive thrombosis of blood with organization on the 
surface. In this form they are difficult to distinguish from organized 
thrombi. In fact Czapek, Thorel and others regard most of the reported 
cardiac myxomas as organized thrombi. The tumors are as small as a pea 
or as large as a hen's egg, and while they usually produce no symptoms 
some seriously obstruct the circulation. They are usually covered with 
blood-clot organized portions of which are fused with the tumor. Fibrin 
emboli may be detached, and Marchand found a growing tumor embolus 
in a cerebral vessel. 

The structure of the tumors is rather uniform but varies as the original 
myxomatous tissue is altered or replaced by edema, hemorrhage, perivascular 
infiltration with round-cells, and organization of secondary thrombi. The 
framework consists of a system of blood-vessels radiating from the pedicle 
which support the overabundant mucinous material composing the bulk 
of the tumor. The vessels are usually thin arterioles or venules or capil- 
laries surrounded by lymphocytes or large mononuclear cells. Muscle- 
fibers may be drawn into the pedicle from the myocardium. An abundance 
of elastic fibers is a common element which Brenner attributes to the mechan- 
ical influence of cardiac contraction upon the tumor stroma. In one of 
Czapek's cases the stroma was nearly cartilaginous. The separation from 
organized thrombi may be based on the highly mucinous character of the 
stroma, on the presence of orderly radiating blood-vessels and elastic fibers, 
and on the absence of masses of blood detritus undergoing organization. 
While many of the larger tumors may be difficult to identify (as in 
Oppenheimer's case) it seems unlikely that a simple organizing blood-clot 
can reproduce the positive features of the true myxoma. Hence the great 
majority of the reported cases of cardiac myxoma are probably genuine. 

The origin of the tumors has been referred to superfluous embryonal 
tissue in the region of the foramen ovale where many of the growths are 
attached. Curtis expressed the view that the process is not a genuine 
neoplasm but a result of chronic inflammation of rheumatic nature producing 
nodular outgrowths of the endocardium which continued to grow because of 
the mechanical influences to which they are exposed. 

Primary sarcomas of the heart are probably connected in origin with the 
myxomas. Binder collected 15 cases, of which eight were located in the right 
auricle. They reach large dimensions, grow into the veins, occlude the heart, 
but rarely give metastases. The composition is chiefly of spindle-cells, 
but giant-cell, and round-cell tumors also occur. 

Apparently pure myxomas have been observed in the mesenterj' (Borst) 
and in the thymus (Winogradow). 

Chondromyxoma may make ud the bulk of the mixed tumor of the 
parotid, and in the fourth recurrence of an original cartilaginous tumor 
of the parotid I have found pure myxoma. Certain of the nasopharyngeal 
polyps which are of cartilaginous origin may appear, especially in recur- 
rences, as pure myxosarcoma. In other organs also, as testis, mammae 
and bones, it seems probable that certain myxomas represent imperfect 
growth of cartilage in embryonal chondromas. 

In several notable cases the intramuscular metastases of cartilaginous 
or choriomatous teratoma testis have appeared in the form of myxoma. In 


fact there is reason to believe that the condition of nutrition of intravascular 
tumors strongly favors the myxomatous type of growth in many tumors. 

The combination of myxoma and lipoma occurs in certain tumors of 
clinical importance and many of these growths reach a large size and prove 
mahgnant. In the retroperitoneal region, in the hilus of the left kidney, and 
in the spermatic cord, are encountered locally malignant myxolipomas. In 
the eleventh recurrence of an original lipoma of the spermatic cord which 
I was able to follow over a period of 15 years some lobules were pure 
myxoma. Robertson, in an analysis of 51 cases of lipoma myxomatodes 
found that 43 per cent, developed in the muscles of the lower extremities, 
33 per cent, in the retroperitoneal region; one reached a weight of 65 lbs., 
and 33 per cent, were malignant. 

In the etiology of myxomas special factors must be considered for different 
groups and an etiological classification becomes of interest. 

A congenital origin has been established for myxomas of the navel which 
probably arise from foci of umbilical cord tissue, and for the mesenteric 
tumor described by Borst. The general source of myxomas must be sought 
in superfluous or aberrant foci of embryonal connective tissue, fat and carti- 
lage. Virchow regarded myxoma as arising indifferently from several 
embryonal mesoblastic tissues, and from glia-tissue. Ribbert would limit 
their origin rather strictly to embryonal connective tissue. The conditions 
which give rise to secondary myxoma, may apparently represent an embry- 
onal reversion of connective tissue, fat or cartilage, or it may result from 
chronic edema and a true mucous degeneration. 

Special embryonic disturbances do not seem to be connected with 
myxomas, although Marchand suggested that abnormalities in the con- 
struction of the cardiac septa might be responsible for myxoma of the 
endocardium. As already stated an intravascular position favors the ap- 
pearance of myxomatous changes in a tumor. Chronic inflammation and 
edema lead to the growth of myxomatoid polyps of mucous membranes. 
Finally in the testis, spermatic cord and possibly in the kidney myxomas 
are probably of teratoid origin, and in some other regions as parotid, breast, 
they represent mixed tumors. 

Sanarelli and Splendore have studied a filterable virus which on inocula- 
tion produces multiple tumors resembling myxoma in the skin and organs 
of rabbits. Rous' tumor of chickens which he was able to reproduce by the 
inoculation of filtered extracts has a myxomatous character. 


A lipoma is a tumor composed of fat tissue. 

The gross appearance of most of these tumors is characteristic. They 
form firm, elastic, rounded, usually multilobulated growths, which, without 
encapsulation, are sharply circumscribed from the surrounding tissue. 
Their size varies from that of a pea to masses weighing many pounds which 
produce serious results from weight and pressure. The consistence is usually 
that of normal fat tissue, lipoma molle, but this density may be reduced by 

Fig. 37. — Lipoma of forearm in adult, containing- fetal fat tissue. 

secondary changes, or, more often increased, by admixture with fibrous 
tissue or by forms of metaplasia. In many situations they become pendulous 
and by constriction of the pedicle spontaneous atrophy may result. The 
color is usually that of normal fat tissue, but xanthomatous changes may 
yield an orange yellow tint, or various secondary processes may give cor- 
responding alterations. Very cellular areas with imperfect fat formation 
are recognizable by a lighter opaque color, while the fully developed tumor 
tissue is yellow and translucent. 




Lipomas exhibit a striking connection with the nutrient blood-vessels, 
each of the multiple lobules growing about a branch of the main vessel. 
Lateral anastomoses of these vessels are scanty so that the tumor grows 
expansively and is readily shelled out of its position. With pure lipomas 
appositional growth is not observed, but this rule is less rigid for mixed 
lipoma and liposarcoma. 

The microscopical structure resembles normal fat tissue, but the lobules 
vary greatly in size and the supporting stroma is irregular in distribution. 
The cells may be overdistended with fat, or may produce the smaller type 
of embryonal fat tissue, and often there are areas of polyhedral cells in which 
the fatty deposits are incomplete. It is from such cells lying in isolated 
foci or along the vessels that the growth of lipomas chiefly occurs. In some 
of these foci an alveolar structure may be produced by polyhedral ceUs 

• ^■. -ji- "d" --.iv - <^f~ ■ 

Fig. 38. — Subcutaneous lipoma. Gross texture of orange tint. Lipoid material appears 

in granular form as in xanthoma. 

with incomplete fat deposits. Occasionally the young cells contain granular 
lipoid pigment and strongly resemble xanthoma cells, and this character 
may be diffuse throughout rather large lipomas, especially about the kidney. 
The blood-vessels are usually overabundant and many cellular arterioles 
are found in the stroma running out into the lobules and dividing into capil- 
laries. Extensive overgrowth of blood-vessels leads to the formation of 
vascular fatty tumors, lipoma telangiectaticum or cavernosum, so that it is 
sometimes difhcult to distinguish between vascular lipoma and angioma. 
Lymph-vessels also may be overdeveloped, and Borst observed the trans- 
formation of a lipoma into a fibrous lymphangioma. 

Secondary changes in lipomas are common in advanced stages of growth. 
One of the more frequent is a mucinous degeneration which occurs in atro- 
phic or edematous areas and may reach extensive proportions without 


constituting a true myxolipoma. DiSuse mucinous degeneration produces 
a peculiar tumor-tissue designated by Miiller as collonema. Mucinous 
changes are more common in liposarcomas, where they usually represent a 
true myxolipoma. Borst has observed the transformation of the fat in 
portions of lipomas into a waxy material. Beriel and Delachanal followed 
the myxomatous and sarcomatous transformation of a lipoma of the sciatic 
sheath which eventually produced lipomatous and sarcomatous metastases 
in lymph-nodes and lung. Calcification either of isolated foci or of the 
entire capsule may occur and is rarely followed by ossification {Lipoma 
petrificum ossificans, Virchow). 

Cysts form in lipomas from the fluidilication of the central portions of 
large lobules producing areas of fluid fat (oil cysts) which are said to result 
chiefly from trauma. In mucinous and edematous lipomas irregular cystic 
areas may form containing mucous or serous fluid, fatty crystals and cal- 
cific deposits. Lymph-cysts with clear fluid are occasionally observed. 

Lipomas may exhibit an excess of fibrous tissue resulting from atrophy 
of fat and fibrous replacement or from admixture with true fibroma (fibro- 
lipoma). More frequently true myxoma is combined with lipoma, and occa- 
sionally chondrolipomas are observed. In a considerable class of tumors all 
the cells are abundant and one has to deal with various types of liposarcoma. 
In the abdominal cavity of a dog I have observed a large tumor composed of 
adult fat, embryonal fat, myxoma, chondroma and large areas of osteoma. 

Many interesting features are occasionally observed in the clinical course 
of lipomas. Some internal growths, as in the thorax, abdomen, or cranium, 
reach such a size and exert such pressure as to cause serious symptoms or 
even death, with symptoms referable to the affected organs. Very large 
lipomas seem to be capable of diverting the nutrition of the body and inducing 
emaciation. This feature is prominent in many of the retroperitoneal 
tumors collected by Voeckler. The limit of growth of simple lipoma is often 
considerable, but in many of the very large tumors portions of the growth 
become more cellular and exhibit sarcomatous structure. The occurrence 
of multiple lipomas, even hundreds (Virchow), is evidence of a peculiar 
dyscrasia of the fat tissues of these subjects, in whom the tumors may appear 
in the lungs and liver where fat is normally absent. A remarkable case is 
that of Broca, in which, in a man of 31 years, after extirpation of a large lipoma 
of thigh, hundreds of small tumors appeared over the body and persisted for 
40 years. The patient finally suffered from dysphagia, with regression of 
most of the tumors, but at autopsy a large fatty tumor surrounded and 
compressed most of the esophagus. Persistent local recurrence of lipomas 
is rare and usually associated with a cellular and vascular or sarcomatous 
structure. Yet in a case of recurrent lipoma of the spermatic cord extir- 
pated many times in the course of 15 years, the structure was never very 
cellular and varied little throughout the course. 

The capacity of the organism to appropriate the fat of lipomas has been 
investigated by Wells, who found no pecuUarity in the chemical composition 
of the fat of lipomas, no deficiency of lipase, and no ground for the common 
assumption that the fat of these tumors is beyond the reach of the fat-mo- 
bilizing factors of the body. Since many lipomas have continued to grow 
while the body was emaciating, and not a few remain unaffected during the 
loss of body fat in phthisis (Madelung), or gastritis (Kuster), it is necessary 
to assume that local conditions, possibly the character of the circulation, 
prevented the absorption of the lipoma. 

Into the etiology of lipoma many factors seem to enter. An hereditary 
influence was recognized in a family observed by Murchison in which the 


father and three daughters had multiple symmetrical lipomas while nine 
sons were free. In Blaschko's case only the male members were affected and 
the tumors appeared at puberty. Other hereditary cases are recorded by 
Meerbeck and Petren. 

The occurrence of multiple symmetrical lipomas has suggested to many 
a connection with the peripheral nerves (Payer, Kottnitz). In a group of 
cases the tumors have been painful and associated with other lesions of the 
nervous system. Alsberg found several neurofibromas with many lipomas 
in the same case and he traced nerve-fibers into certain lipomas. Both 
xanthoma and multiple lipoma have also been observed with multiple neuro- 
fibroma. Yet the occurrence of lipomas with multiple neurofibroma is 
extremely rare and it has not been possible to estabhsh for lipoma such a 
relation to the peripheral nerves as exists with fibroma. Grosch and others 
have argued that multiple lipomas may originate in connection with the se- 
baceous glands, as the result of a trophoneurosis, but later observers have 
been unable to find support for this theory (Lit. Goebel). 

Virchow described as capsular lipoma certain limited but sometimes bulky 
growths occurring in atrophying organs as the kidney and breast. A similar 
growth of fat tissue replaces the atrophic thymus and bone-marrow and is 
very often seen in the atrophic lymph-nodes of cachectic or emaciated sub- 
jects and with chronic mastitis. Askanazy was able to show that in several 
cases of small multiple lipomas the tumors were located in atrophic lymph- 
nodes, and he suggested that this was a common origin of lipomas. The term 
replacement lipomatosis seems more applicable to this process which may 
produce fatty growths of moderate dimensions but not true lipomas. 

Multiple lipomas have been observed in many cases of disease or atrophy 
of the thyroid gland in cases of obesity and in alcoholism, but it does not appear 
that these constitutional conditions have any direct bearing on the etiology 
of true lipoma (Curling, Madelung, Kottnitz). They seem more clearly 
connected with the forms of diffuse or regional overgrowth of fat tissue which 
bear the same relation to lipoma as diffuse fibromatosis holds to fibroma. 
In such conditions disturbances of the thyroid and pituitary glands seem to be 
an important factor. 

A congenital tissue predisposition seems to be an essential facor in the 
origin of most lipomas. This predisposition may take the form of a dis- 
turbance of development of the fat tissue. It has been shown by KoUiker 
and Toldt that the fat lobules have a certain independence in development 
which constitute them a sort of primitive organ. The comparative isola- 
tion of the blood-supply of these lobules reappearing in lipomas suggests 
that the anomaly predisposing to lipoma is connected with the distribution of 
blood-vessels in the fattv tissues. 

A more definite embryonic disturbance is probably responsible for the 
origin of many lipomas of the internal organs, as in the kidney, where mis- 
placed islands of capsular fat are held to give rise to tumors (Selter, Lu- 
barsch, Manasse, Miiller). 

More complex embryogenic anomalies are concerned with those lipomas 
which arise in the cranial cavity with cholesteatoma, where complex epidermal 
tissues are involved (Bostroem); with spina bifida (v. Recklinghausen, 
Arnold); with a rudimentary cervical rib (Volcker); in the uterus (Merckel). 

Certain lipomas result from predominance of fat tissues in mixed tumors. 
Such is probably the nature of the recurrent lipoma of the spermatic cord. 
(Sazarin, Ehrendorfer, Forges). 

Trauma of many types has acted as an exciting factor with many soli- 


tary superficial lipomas in subjects in which a local or general predisposition 
must be assumed to exist. 

In the general etiological classification of lipoma and lipomatoid processes 
one must recognize many different forms of hypertrophy of fat tissue as fol- 
lows: (i) Obesity; (2) locahzed overgrowth of fat tissue, lipoma annulare colli; 
(3) replacement lipomatosis, as in atrophic organs, marrow, kidney, capsular 
lipoma, lymph-nodes; (4) homologous lipoma, a group including the ma- 
jority of sohtary subcutaneous lipomas; (5) heterologous lipoma from mis- 
placed groups of embryonal tissue cells; (6) overgrowth of lipoma in mixed 
tumors and teratomas. 

Clinical Types of Lipoma. — Subcutaneous. — The subcutaneous tissue is the 
commonest seat of hpoma, and the back, neck, shoulders, axilla, and abdom- 
inal wall are most frequently affected. According to Grosch who has plated 
a large number of subcutaneous lipomas, these tumors occur in inverse 
proportion to the number of glandular structures in the skin. The cutaneous 
lipomas are single or multiple, unilateral or bilateral and symmetrical, 
small or large, and polypoid, or pendulated. They arise in the derma, 
beneath the superficial fascia and in the deeper fascias. 

Small congenital lipomas occur beHeath the skin on the volar side of the 
fingers and hands. They arise from the deep fascia and may be connected 
with the tendon sheaths. 

Lipoma annulare colli is a diffuse form of Hpomatosis occurring in the neck, 
producing great enlargement of this region, and involving the fat tissues of 
the skin and subcutaneous and intramuscular structures (Madelung). It 
is not a true lipoma but a locahzed overgrowth of the abundant fat tissues of 
this region. Somewhat similar overgrowths of fat tissue may occur in the 
hips and thighs (Shattock). More general hypertrophy of fat tissues is 
seen in adiposis dolorosa (Dercum). Subcutaneous lipomas often reaching 
considerable size occur in the scrotum and labia. 

Synovial. — Lipoma and lipomatous processes affect the joints. Soli- 
tary lipoma may arise within the joint cavity through rupture of the syn- 
ovial membrane and protrusion of the subcapsular fat which then goes on to 
continuous hyperplasia (Konig). Solitary lipomas may also arise in the 
joint from overgrowth of fat in synovial fringes (Otterbeck, Filter.) Ex- 
tra-articular lipomas arising from the subcapsular fat occur about the knee 
and hip. Lipoma arborescens is a characteristic extensive papillary outgrowth 
from the fat tissues of the synovial membrane and joint capsule of the knee 
which fills the joint spaces and interferes with function. It occurs with 
simple chronic and with tuberculous arthritis in which case it is probably 
of inflammatory origin with overgrowth of synovial fat (Kaufmann), but it 
occurs apart from any inflammatory processes as a true tumor of pure fat 
tissue in which the characteristic relation of lobules to branching subcap- 
sular blood-vessels is observed (Borst, Ribbert). 

Intermuscular. — In the cheek a congenital lipoma arises from a mass of 
fat in the canine fossa, on inner surface of masseter muscle, long known as 
corpus adiposum mala (Bichat), pushes its way beneath the skin and appears 
as a subcutaneous tumor of moderate dimensions. Several tumors of this 
origin have been collected by Bruns and Ransohoff. They may readily 
be mistaken for tumors of the parotid with which they are in intimate 

Beneath the pectoralis hpomas arise probably from extensions of the 
axillary fat which cause protrusion of the breast (Billroth). Deep lipomas 
are also observed in the orbit, at the base of the tongue, and in the larynx. 
Grosch has collected many Hpomas occurring about the cranium. 


Cranial. — In the cranial cavity lipomas have been observed in various 
situations in most of which small collections of fat tissue are normally pres- 
ent. The chief locations include the surface of the corpus callosum, the base 
ofjthe cerebrum, the brain stem and cerebellum, the ventricles, and the roots 
of the cranial nerves (v. Sury, Lit.). They invariably arise from the pia 
although many appear to be imbedded in the brain tissue. Taubner 
argued that they may arise from the glia-tissue. 

Virchow mentions six lipomas of the pia, one of the raphe of the corpus 
callosum, and four at the base. In the choroid plexus and nourished by its 
vessels they have been observed by Virchow, Wallmann and Hackel. Rind- 
fleisch saw small multiple lipomas of dura and ventricular ependyma. Fere 
and Francillon report symmetrical lipomas. Garnier's lipoma was attached 
to the crus cerebri. Small lipomas of the corpora quadrigemina are reported 
by Bernhard and Taubner. The most characteristic type is the lipoma which 
is nourished by the vessels of the corpus callosum, grows along this structure 
and largely replaces the nerve tissue (Virchow, Benjamin, Coats). Peri- 
neural hpomas inclose the roots of the cranial nerves or olfactory bulbs 
(Shouppe). A central nucleus of bone has been found by Benjamin, Chiari, 
Ernst, and v. Sury. 

Many of the tumors are small and fail to produce symptoms, while not 
a few reach sufficient bulk to cause atrophy of adjacent nerve tissue. Weil 
removed at operation a lipoma, lo X lo X 4 cm. from the temporal fossa. 
Wurth's lipoma of the basal pia extended into the left hemisphere and caused 
epilepsy and hemiplegia. From the studies of Bostroem who collected 28 
cases it appears probable that most intracranial lipomas originate from epi- 
dermal inclusions although no dermal elements have been found in them. 
Nippe found a lipoma of the parietal lobe surrounded by gliosarcoma which 
he attributed to trauma. 

Lipomas may develop in connection with defective development of the 
spinal canal and meninges. . A series of tumors, some of which were pure 
hpoma, others composed of fat, fibrous tissue or cartilage, have been observed 
in connection with spina bifida and with scantily developed meningocele 
(v. Recklinghausen). Arnold's hairy fibrolipoma of the scalp appears to 
have originated from imperfect closure of a cranial suture. 

Renal. — In the kidney small circumscribed multiple lipomas occur be- 
neath the capsule or replacing a portion of cortex, or reaching to the medulla. 
The structure of these tumors varies. Some of them are composed of pure 
fat tissue; others contain considerable fibromatous tissue, so that Virchow 
and others have spoken of them as fibrolipoma. In either case the light 
yellow translucent color seems to distinguish them from the more common 
struma suprarenalis of Grawitz, which is opaque and of an orange tint. 
In not a few cases smooth muscle-tissue has been found in such tumors and 
Selter, Lubarsch, and Manasse have described them as myofibroma; while 
in other cases sarcomatous features are observed as in the case of Miiller's 
(myoHposarcoma). I have observed a very large liposarcoma of this 
type, in one lobe of which xanthomatous changes were pronounced. Borst 
and Selter describe symmetrical lipomas of the kidneys chiefly of the upper 
half of the organ. In Borst's case double myolipoma was associated with 
fibro-muscular tumors of the same portions of both kidneys. All these 
features point to a congenital and embryonal origin of these tumors, but the 
exact nature of the embryogenic disturbance is undetermined. That they 
may be connected in some way with Grawitz' struma suprarenalis is indicated 
by the presence of adrenal tissue in Mtiller's case, and by the occurrence of 
curious, complex, chiefly liposarcomatous tumors of the kidney, in one of which 


I have observed areas of perithelioma mingled with liposarcomatous tissue. 
Somewhat similar myolipomatous growths occur along the spermatic cord. 

Partial or complete replacement lipomatosis of the kidney has been ob- 
served in a series of cases reported by Lacrampe-Loustan. Rayer and Ep- 
stein have described such cases in which no trace of renal tissue remained while 
the intact capsule of the organ inclosed pure fat tissue. Seller called atten- 
tion to the fact that in most of these cases a single large calculus occupies the 
renal pelvis. The proliferation of fat begins in the pelvic adipose tissue, 
and follows atrophy of the renal parenchyma. 

True pelvic lipomas probably develop from such capsular tumors which 
early project into the cavity. Warthin described a very large peduncu- 
lated intrapelvic growth over which the mucosa and atrophic kidney tissue 
was widely stretched, while a tongue-hke mass extended several inches into 
the ureter. 

Perirenal lipomas containing connective tissue, mucoid areas, and often 
sarcomatous areas occur in infants and adults and may reach very large di- 
mensions. In advanced stages they pass as retroperitoneal lipomas. Lau- 
wers describes a tumor arising at birth and reaching a weight of 6 lb. at seven 
years. Windle's liposarcoma weighed 50 lb. Adami described two very 
large tumors composed of adult fat tissue with cellular stroma. The growth 
is usually slow but with sarcomatous structure the progress may be rapid and 
recurrence is frequent. 

Gastrointestinal. — In the gastrointestinal tract lipomas are of rare occur- 
rence. They arise from the submucosa and from the appendices epiploicas 
of the colon. Dewis collected 44 cases in nine of which the tumor was expelled 
spontaneously while intussusception occurred in 21. The subserous lipomas 
m.ay be worked loose in the peritoneum. Ehrlich has collected 52 cases of 
intestinal lipomas. In the gastric submucosa small fatty tumors may occur. 
I have observed a large fatty tumor 4X 10 cm. surrounding the appendix 
which was the seat of chronic suppuration. Very large retroperitoneal 
lipomas have been observed. (Madelung, Waldeyer, Adami.) 

Mediastinal. — In the mediastinum localized overgrowth of fat tissue is 
not unknown, especially in alcohohc and obese subjects. True lipomas of 
the mediastinum may reach large dimensions and cause pressure symptoms 
from the thoracic organs. Arising from many different points in the sub- 
pleural fat, mediastinal lipomas may project into the pleural cavity or along 
the intercostal spaces (Rokitansky). Or the growth may perforate the 
chest wall and appear externally in the back (Czerny), beneath the breast 
(Gussenbauer) or on the front of the chest (Cruveilhier, Conner). In the 
writer's case the tumor encircled nearly all the structures in the thorax. 
In Fitz's case there was congenital hypoplasia of the left lung. Carless 
observed a large lipoma of the thyroid region of which the pedicle was traced 
into the mediastinum and which was associated with other multiple symmet- 
rical lipomas. I have seen one mediastinal liposarcoma with tumor fragments 
in sputum and metastasis in deltoid muscle. 

Cardiac. — In the heart lipomas have been described by Ribbert, Orth, 
Petrocchi, Spalty, Hagedorn and others. They may originate in misplaced 
islands of epicardial fat tissue. Dittrich's tumor was congenital. 

Uterine. — In the uterus Orth has observed a polypoid lipoma and Wer- 
kel two intramural lipomas, while lipomatous areas of mixed tumors of the 
uterus are described by Gebhart. Pollak thought his intrauterine lipoma 
originated from a portion of omentum protruding into a wound of the uterus. 

Myelogenous. — A true hpoma of the marrow of the fibula in a young girl 
has been recorded by Wehrsig. 


Localized overgrowth of cartilage occurs in several forms between which 
it is sometimes difficult to draw sharp distinctions. Limited outgrowths 
of preexisting cartilage occur on the ribs, in the larynx, and about joints, 
which exhibit the characters of a simple hyperplastic process, and are called 
ecckondroses. True progressive neoplasms composed of cartilage appear 
in the same situations and also in tissues not normally containing cartilage 
and these are called chondromas or enchondromas. That many ecchondroses 
possess some of the properties of tumors is indicated by the occasional 
occurrence of large chondromas as a sequel of ecchondrosis. Thus chondro- 
mas as large as an apple occur at the chondrosternal junction. In a case 
of Weber's there was striking symmetry in the location of multiple chondro- 
mas and ecchondroses. 

Virchow classed as ecchondroses the smooth difiuse outgrowths of 
permanent cartilage, and as enchondromas the large circumscribed lobulated 
tumors of transitory cartilage. Between these main classes are many 
intermediate forms which display more or less clearly the neoplastic quahties. 

The ecchondroses are usually small, multiple, smooth and nodular or 
diffuse outgrowths of preexisting cartilage. After reaching a certain stage 
their growth tends to culminate, often with the natural growth of cartilage 
in the body, after which they may remain quiescent or undergo secondary 
changes, chiefly calcification or ossification. Amyloid deposits may also 

In structure the ecchondroses copy normal hyaUne and fibrocartilage 
according to their points of origin, and show few of the atjrpical features 
of true chondroma. They are usually surrounded by normal perichondrium 
from a portion of which, according to v. Rindfleisch, they originate. 

Ecchondroses are not extremely common but rather widely distributed. 
The most familiar examples occur at the costochondral junctions where 
they appear as smooth multiple nodular swelhngs which tend to calcify or 
ossify. Virchow and others believed that these and most other ecchondroses 
were a sequel of rickets and this origin probably applies to the ecchondroses 
arising in early life and soon tending to calcify. 

A well-known form of ecchondrosis appears on the inner surface of the 
symphysis pubis, and may reach such dimensions as to obstruct labor 
(pelvis spinosa) . About the epiphyses of any of the long bones small nodular 
outgrowths of cartilage may appear. 

Localized cartilaginous outgrowths of the cricoid and thyroid cartilages 
have been described by Virchow and many later writers. They project 
inward with deformation of the larynx but with rare exceptions do not 
obstruct breathing. The entire cricoid cartilage may be diffusely thickened 
by multiple ecchondroses. From the upper and lower surfaces of the tracheal 
rings multiple cartilaginous nodules may develop. They are connected 
with the rings by strands of fibrous tissue or perichondrium, and they may 
.become so numerous as to cause fusion of many adjacent rings (Virchow, 
Recklinghausen, Mischaikoff). Ecchondroses usually of small size develop 
from the intervertebral discs and project outward or into the spinal canal. 



In the joint cavities multiple ecchondroses are a frequent result of chronic 
arthritis with osteochondritis. The modes of origin of these growths and 
the manner in which they become detached, forming joint mice, have been 
exhaustively discussed by Virchow. 

Enchondroma. — The true chondroma is a rather common tumor, the 
varied features of which, perhaps more than any other benign tumor, il- 
lustrate the pecuharities of neoplasms. This result is perhaps dependent 
upon the facts that cartilage is essentially an embryonal and transitory tissue, 
and that cartilage cells although encased in a firm matrix have rather active 
proliferative powers, possess ameboid properties and are readily subject 
to metaplastic changes. For the most striking display of these characters 
one must pass to the tumors of lower animals with which chondromas are 
more common than with man. 

The chondroma produces a hard, rounded, lobulated tumor which often 
reaches large dimensions. The lobulated structure is referred to the expan- 
sive growth from multiple centers, as a result of which the surrounding tissues 
are pressed aside. In this way large tumors may form, composed of con- 
voluted masses resembling the convolutions of the brain. Such growths are 
commonly encapsulated by perichondrium and are cartilaginous throughout 
at all stages. Other chondromas are imperfectly encapsulated, encroach 
upon surrounding tissues as a malignant tumor, and invade blood-vessels. 
In such cases one finds many stages of the formation of the perfected cartilage 
from more cellular tissue, the nature of which is not always clear. The 
nutrition of the solid chondromas is maintained by a system of lymph 
spaces from the periphery. Rindfleisch and Borst have traced a rich system 
of lymphatics from the periphery of a submaxillary chondroma and com- 
municating with the pericellular spaces. Through most of these growths 
there are irregular bands of connective tissue carrying blood-vessels and this 
feature becomes pronounced as the tumors tend to calcify or ossify. About 
many chrondromas there is an excessive development of blood-vessels to 
which Virchow and v. Recklinghausen ascribe an important and primary 
influence in the development of the tumors. Such growths have been termed 
angiochondroma and a few of them have been deeply pigmented by diffusion 
of extra vasa ted blood pigment (Siegert). 

The capsule of subperiosteal chondromas may consist chiefly of the 
thickened periosteum which may go on to produce bone, and a bony capsule 
may surround chondromas developing in the marrow cavities. 

While most chondromas are solid, many become softened by mucinous 
degeneration, and cysts filled with mucinous, serous or fatty material may 
form. The skin over a chondroma may atrophy, and ulceration, excavation 
of portions of the tumor, and suppuration may be established. Although 
the chondromas are usually localized and benign they sometimes grow 
extensively in the blood-vessels, filling the lumen with nodular or solid masses 
and extending over wide areas. In a case reported by Ernst a chondroma 
of the lumbar spine invaded many of the abdominal and pelvic veins, traveled 
up the vena azygos and vena cava, filled the right ventricle and continued 
into the pulmonary artery. Discontinuous metastases also occur especially 
in the lungs, occasionally in the lymph-nodes. 

The multiple character of certain chondromas has been a striking feature, 
and in some cases nearly every bone in the body has been affected. In 
the notable case reported by Recklinghausen, the hands, feet, knee, elbow, 
and ribs were the seat of very numerous chondromas. 

The location of skeletal chondromas is usually in the diaphyses of the 
bones near the epiphyses. They may project externally, pushing outward 


the periosteum, or growing in the medullary cavity they distend the cavity 
and are surrounded by the thin bony shaft. 

The structure of chondromas reproduces that of the various normal 
types of cartilage, chiefly the hyahne variety. Ranvier classified chon- 
dromas, according to their structure, in four groups, (i) A single lobe of 
hyaline cartilage. (2) Several lobules of hyaline cartilage separated by 
fibrocartilage. (3) Fetal cartilage. (4) Cartilage with stellate cells. 
The cells maj' be more or less numerous than in normal cartilage. They 
vary greatly in size and usually lack the orderly arrangement into groups of 
cells with opposed surfaces flattened. Peculiar vacuoles are common in 
these cells and were described by Virchow under the term "physaliden. " 
The cells are usually rounded and lie in distinct spaces, but with imperfect 
formation of chondrin matrix they become stellate with mucinous degenera- 
tion, fusiform as they approach the connective tissue type, calcific granules 
may be deposited in them, and in some cases they show all transitions up to 
the bone cell in osteoid and osseous areas. The cells usually contain glycogen 
and fat granules and well-marked fatty degeneration may occur. 

The growing cells of a chondroma are not as a rule those inclosed in the 
matrix but rather those on the periphery of the tumor. The matrix of the 
chondroma is usually hyaline, but may be fibrillated or may contain elastic 
fibrils. Ernst noted very active production of the elastic fibrils in a rapidly 
growing tumor. The matrix is not an integral part of the cells but a product 
which is deposited under cellular influence (Ribbert). The chondrin is 
usually deficient and irregularly distributed and various tj^es of secondary 
change are frequent. 

Calcification with the deposit of phosphate and carbonate of lime affects 
first the matrix, producing an irregular network of densely basic staining 
material which is composed of granules compacted together and becoming 
homogeneous. The borders of the cell spaces may be first affected and from 
this point the deposit appears in the cells so that in advanced stages both 
matrix and cells may become heavily incrusted. In such areas the cells 
assume a stellate form approaching the character of bone cells. 

Ossification of chondromas occurs under several conditions. An im- 
perfect formation of islands of bone may take place at different points 
throughout chondroma, this change being preceded by increased vascularity 
of the septa, and by calcification and final ossification. 

More often the connective tissue in and about the chondroma having 
first laid down cartilage goes on to replace it by trabeculae of well-formed 
bone surrounded by typical osteoblasts. In this process the normal events 
in the production of bone in cartilaginous matrix take place. 

About central chondromas which invade the shaft of a bone the peri- 
osteum may be incited to the production of a bony capsule which is of 
inflammatory origin and not a part of the tumor. 

In many chondromas the formation of cartilaginous matrix is imperfect 
and a simple hyahne or partly fibrillated or mucinous material takes its 
place. This imperfect chondrification is much more common than any 
form of softening of chondrified matrix. In the soft matrix the cells may 
attain large size and he in very large round spaces, or, in myxochondroma 
they become stellate and the matrix fibrillated. Some of the soft chondromas 
are composed of a shell of perichondrium from which spring papillary fringes 
of poorly chondrified cartilage while the central portions are filled with 
mucinous fluid (cystic, papillary chondroma). 

Distinctly malignant cartilaginous tumors, chondrosarcomas, occur in 
several forms. Medullary chondromas may perforate the shafts of long 



bones and destroy the epiphyses, proving mahgnant from mechanical fac- 
tors. Infiltrating chondromas in which very cellular peripheral portions 
invade surrounding tissues and produce nodules of cartilage are properly 
called chondrosarcoma. There is a considerable group of chondromas occur- 
ring in mixed tumors and teratomas, which are usually very cellular, 
containing other tissues besides cartilage, and which invade tissues, lymph- 
nodes, and blood- and lymph-vessels. Invasion of veins is not confined to 
the embryonal or teratoid chondromas, for in Virchow's case of tumor of 
fibula and in Ernst's case of fibrochondroma of spine there were extensive 
intravascular growths, while the pelvic chondromas in several cases have in- 
vaded blood-vessels and lymphatics (Foerster, Weber, Biesiadecki). In 

Fig. 39. — Structure of the wall of a cystic chondroma. 

this group the malignant properties are often out of proportion to the cel- 
lular quahty. In a group of sarcomas of long bones composed of cartilage, 
bone, and cellular tissue, the local destructive effects and recurrence after 
operation lead to their designation as osteochondrosarcoma. 

While in general the more cellular and softer tumors, as myxochondroma, 
are the more malignant, both chnical and histological features should be 
regarded in applying the term chondrosarcoma. The metastases of chon- 
dromas exhibit most of the characters observed with other mahgnant tu- 
mors and a special capacity to invade the large veins must be ascribed to 
them. Invasion of the lymphatics also occurs. _ The structure of^ these 
mvasive chondromas does not always suggest malignancy. In Ernst's case 
the intravascular cartilage was provided with firm hyaUne matrix, but this 



tumor did not produce embolic nodules. To account for the invasive prop- 
erties one must assume that the active growth and infiltrative properties 
reside in the undifferentiated cells of the periphery of the cartilaginous'nod- 
ules and in some cases extensive continuous growths in vessels and emboKc 
metastases have consisted almost exclusively of such undifferentiated cells. 
By such changes a chondrosarcoma may lose much or possibly all of its chon- 
dromatous character. This possibility also suggests that certain tumors 
originating from the mother tissue of cartilage may never show pronounced 
chondromatous features but appear as simple sarcoma or myxosarcoma. 
Certain nasopharyngeal sarcomas (polyps) are probably of this character. 

Etiology. — The majority of chondromas occur in early life and about 
puberty. Some are congenital, and multiple chondroma of the spine is 
hereditary and congenital. Many of them are located at the growing ends 
of bones and in not a few there is a distinct tendency toward standstill and 
regression at the period when the development of the skeleton is complete. 
All of these facts point to an origin from disturbances in development of the 

Fig. 40. — Island of superfluous cartilage in periosteum of rib. (£. Milller.) 

cartilaginous elements in the formation of bones and joints. An inherited 
predisposition hag been emphasized by Weber, who observed many chon- 
dromas and ecchondroses in several members and three generations of the 
same family. In the French family of Pellerin members of three generations 
were affected with multiple chondromas of tibia, ribs, and humerus (Ernst). 

Virchow, Recklinghausen and many others have shown that rickets is an 
important factor in the disturbance of the growth of bones that leads to 
chondroma. Borst has found widely misplaced islands of cartilage in ra- 
chitic bones and believes that these may give origin to chondroma of limited 
growth, but doubts the general applicability of the rachitic theory of skele- 
tal chondroma. Ribbert also fails to find in this theory a sufficient ex- 
planation of any large group of chondromas and thinks the disturbance must 
result from some unknown fetal disorder of the bones. E. Miiller has de- 
scribed superfluous islands of cartilage in the periosteum of the ribs and sug- 
gested that these structures may be the source of chondromas. 

In the remarkable case of Recklinghausen's in which multiple chondromas 
of feet and hands were associated with angioma, the author assumed that 
congenital aplasia of the blood-vessels subsequently leading to passive dila- 


lation and angioma, prevented the normal progress of replacement of car- 
tilage by bone and resulted in overgrowth of cellular cartilage. The tumors 
ceased to grow at 22 years. 

For the large group of distinctly heterologous chondromas of breast, pa- 
rotid, neck, uterus, etc., various embryonic disturbances must be assumed to 
exist, in most of which complex masses of tissue are involved. Chondroma 
of the sex glands results from the one-sided development of teratomas and 
it is probable that in other localities certain chondromas have a similar origin. 

Trauma and inflammation are frequent causes of ecchondroses, and 
trauma is not infrequently reported as preceding chondroma and chondrosar- 
coma (50 per cent. Weber), especially of the long bones, in which cases it 
must be assumed that there is also a local or general predisposition. 

Clinical Types of Chondroma. — The skeletal chondromas are the most 
frequent forms of this tumor. They occur most often as single or multiple 
tumors of the hands and feet but may affect any of the bones (Nasse, Lit.). 
They arise from periosteum or in the medulla, usually near the diaphysis 
and regularly before puberty, tending to become stationary at that period. 
They deform the shafts of the bones and the joints. From the pelvic 
synchondroses they may grow to very large size, a case of Weber's having 
reached a diameter of two feet. These tumors as well as the pubic ecchon- 
droses may obstruct labor. In a girl of 15 years Pfeiffer observed a pelvic 
chondroma which measured one meter in circumference and weighed 27 
pounds. Most of these large pelvic tumors have shown rapid growth 

Pronounced cases of multiple osteochondroma represent a somewhat 
specific malady which is probably related to chondrodysplasia and other 
disturbances in the growth of bone (Ashurst, Lit.). It is responsible for a 
variety of skeletal deformities, as depicted by Hagen. The disease some- 
times exhibits hereditary tendencies, being transmitted by affected males 
and females and by unaffected females. The tumors may appear at birth, 
or be delayed until the soth year, and one tumor may be growing while 
another is regressing. There is generally a hmit of growth about puberty. 
The growth of the long bones may be so inhibited that the subject is dwarfed. 
Growth of radius over that of ulna may produce a deformity resembling 
dislocation of the radius, and growth of tibia over fibula results in pes valgus. 
The x-ray shows osteoporosis of the ends of the bones, and often a cystic 
appearance, while the compact bone of the ends of shafts may be very 
deficient. At various points, usually about the joints, the multiple 
outgrowths appear. The structure shows a persistence and overgrowth of 
poorly ossified or calcified cartilage, in which the cells are irregular in size 
and form (Carman, Fisher). The ordinary epiphyseal fine is irregular or 
obhterated. Late ossification leaves the bone deformed and covered with 
osteophytes. Lenormant collected a series of cases in which osteosarcoma 
appeared to develop on the basis of the early lesions. 

Chondroma of the scapula arises from various portions of the bone, and 
exhibits the structure of the benign chondroma or that of chondrosarcoma, 
often myxomatous. The great variations in the structure and course of these 
tumors is fully presented by Deganello in a study of 39 cases. Osteoid 
chondroma is a pecuhar form of skeletal chondroma arising beneath the 
periosteum of the long bones. As the growth enlarges the bone assumes a 
thickened spindle form. The structure is that of hyaline acidophiHc osteoid 
tissue in which the cells approach the form of bone-cells, and true bone may 
form in them. Many of the osteosarcomas correspond to this description. 



Virchow describes an extensive case, largely ossified, arising in the soft 
intercostal tissues and producing metastases in the lung and pleura. 

Branchiogenic chondromas occur in the neck from misplaced islands of 
cartilage derived from the branchial clefts. They are located in and about 
the sternomastoid ligament and muscle with branchiogenic cysts and fistulas, 

Fig. 41. Fig. 42. 

Multiple congenital ecchondroses in a negro girl of ii years 

Fig. 41 
A. S.. 63.) 

Fig. 42. — Multiple congenital osteochondroma. 
(After Carman and Fisher, A. S., 6l.) 

(After Ashurst, 
X-ray photograph of bones of forearm. 

and near the parotid gland. Zahn has collected twelve cases in which mis- 
placed islands of cartilage were discovered in this region, and Bidder observed 
a congenital tumor arising from such an island. Recklinghausen's chon- 
droma of the thyroid may have arisen from such a tissue rest. Symmetrical 



islands of cartilage and bone have been found in the tonsil but do not seem 
to have given rise to tumors (Deichert). 

_ An auricular chondroma is described by Virchow as springing from a 
misplaced portion of the fibro-elastic cartilage of the ear. 

The chondromas of the hyoid region described by Boeckel and Spisharny 
seem to have been connected entirely with abnormalities of the hyoid bone'. 
The various embryological disturbances giving rise to cysts, rests, and tumors 
in this region will be discussed elsewhere. 

Chondroma of the respiratory tract occurs in the trachea, bronchi, and 
lungs, apparently arising from the same structures that lead to ecchondroses. 
Virchow has described multiple chondromas of the lung, located at the root, 
in the parenchyma and on the pleura. While often purely cartilaginous, they 

Fig. 43. — Structure of multiple osteochondroma. Section through outer bony shell. A, 
Hyaline cartilage with atypical cells and calcium deposit. B, bone trabecule. C, mucoid 
connective and fat tissue. {After Carman and Fisher.) 

may become calcified or ossified. In the lung rare tumors occur as multi- 
lobed growths of hyaline cartilage, subpleural or deep within the lobe, and 
their association with connective and fat tissue points to an origin for some 
of them from a complex mass of tissue. In Siegert's case the chondroma 
projected from the bronchial wall into a dilated bronchus. Chronic thicken- 
ing of the pleura may yield flat plates resembling cartilage but composed of 
hyaline or calcified connective tissue. Yet Kramer has described a lobulated 
chondroma of the pleural surface of the diaphragm. 

In the breast pure chondroma is occasionally observed but more often the 
cartilaginous growth is a portion of a mixed tumor usually of epithelial 

It is commonly assumed that in all instances these tumors arise from 
isolated masses of tissue which include a skeletal portion giving rise to the 
chondroma, but the correctness of this view is not demonstrated. It is 
possible that some of the cartilaginous tumors arise by metaplasia of other 


tissues. In the dog a large proportion of mammary cancers contain cartilage 
and the histological structure strongly suggests its formation by metaplasia 
from other tissues. WilUams has collected many reports of cartilaginous 
and bony tumors of the breast and argues in favor of the origin from mis- 
placed masses of complex tissue. 

In the uterus, especially in the cervix, cartilaginous and bony tumors 
are occasionally observed (WiUiams, Lit.). It seems probable that they arise 
from misplaced islands containing skeletal tissue. Thus Meyer found a 
bony nodule in the cervix associated with a remnant of the Wolffian duct. 
Miller found an osseous tumor replacing the corpus in a subject in whom 
the vagina was absent. In several uterine myomas cartilage has been found 

In the mixed tumors of the salivary glands cartilaginous areas are very 
common, and especially in the submaxillary the tumor may be nearly pure 
chondroma. Here the cartilage is of imperfect hyahne variety and often 
of myxomatous type. In recurrences of these tumors there may be a tend- 
ency toward elimination of all but the cartilaginous elements. The various 
questions that concern these tumors will be discussed elsewhere. 

Teratoid chondroma. A considerable group of chondromas, including 
many of those which invade the vessels and prove mahgnant, represent 
one-sided development of the cartilaginous portions of teratomas. Most 
of these chondromas arise in the sex glands, ovary and more especially the 
testis, but with the extensions of the limits of occurrence of teratomas it 
seems probable that chondromas of this origin may be more numerous than 
now appears. In the testis these tumors are sohd or cystic, and other 
elements besides cartilage are usually present in abundance. The cartilage 
may be reduced to very minute traces which escape detection in many tumors 
or tridermal type (Ewing). In several notable cases extensive intravascular 
growths and metastases occurred (Paget, Ohkubo, Lit.). 

Chordoma. — A remnant of the chorda dorsalis, a specific entodermal 
embryonal tissue about which the spinal column develops, regularly persists 
in infants in the centers of the intervertebral discs (Kolliker). It was 
shown by H. Miiller that portions of chorda tissue commonly persist at the 
base of the skull and in the coccyx, and that this tissue probably gives 
origin to certain peculiar tumors which Virchow first described as ecchondro- 
sis spheno-occipitalis. Ribbert found chorda remnants in the bony tissue and 
flat plates beneath the dura mater of the Chvus Blumenbachii, and held 
that they produced tumors only when misplaced and lying outside the bony 
encasement. He found the tumors in about 2 per cent, of a series of 
subjects, in the form of bean-shaped masses which perforate the dura and 
become adherent to the basilar artery with which they may be torn away in 
removing the brain. By careful sectioning he was able to show that the 
small intradural tumors are connected by a pedicle with the underljdng 
bone in which their roots are imbedded. They seldom reach a large size 
and being soft do not produce pressure or other symptoms. Yet Grahl 
observed a lobulated extradural chordoma of the sella turcica which produced 
fatal pressure symptoms. A large chordoma lying in front of the upper 
cervical vertebrae has been observed by Klebs. Hassner's tumor measured 
5.5 X 6.5 cm. and had produced pressure symptoms for four years. 

In structure chorda tissue is composed of groups of large vacuolated 
cells lying in soft homogeneous basic staining matrix, and the tumors accu- 
rately reproduce this structure. Virchow described the pecuUar vacuolated 
cells of these growths under the term ecchondrosis physahfora, and since some- 
what similar vacuolation occurs in chondroma, he was inchned to regard 


all of them as chondromas. Although chorda tissue does not produce 
cartilage and there can be no transition forms between chondroma and chor- 
doma, yet it may at times be difficult to distinguish between them. 

Malignant Chordoma. — The exact limits of the occurrence and possible 
variations in structure and course of chordoma have not been determined. 
It would seem likely that a tissue as extensive as the chorda dorsalis might 
persist at various points along its normal location, and that it might give 
origin to structures varying from that of the simple chordoma. 

Nebelthau has shown that remnants of chorda tissue may occur at several 
points along the base of the skull. Fischer has described a malignant chor- 
doma arising beneath the dura of the clivus B., invading the spinal canal and 
compressing the cord. A small vein was invaded by polyhedral tumor-cells. 

1 * 


• '••: 


• .* v« .-., ».. . ,*-'2/'* . •? 

Fig. 44. — Sacral chordoma. (From a section received from Dr. Mallory.) 

The structure corresponded to that of chordoma but some areas resembled 

Link obtained portions of a tumor in the pharyngeal vault, which affected 
the internal ear and compressed several cranial nerves and which he believed 
arose from the clivus B. Besides the usual chorda-like tissue it contained 
groups of polyhedral tumor-cells so that Linck recalls the entodermal origin 
of the chorda and discusses the question whether the malignant chordoma 
should be called sarcoma or carcinoma. Coccygeal chordomas are described 
by Feldmann and Vecchi, and in the latter case the growth recurred after 
operation. Albert describes a coccygeal chordoma which involved the 
rectal wall and recurred promptly after operation. 

The positive identification of chordoma is not readily accomplished. Two 


common tumors closely simulate or even duplicate the structure attributed 
to chordoma — viz., myxochondroma and colloid carcinoma of the intestinal 
canal. The gross relations of the tumor should be given first consideration in 
the diagnosis. Especially in the sacral region chordoma is difficult to separate 
satisfactorily from chondroma or colloid carcinoma. Feldman analyzes in 
detail the features which distinguished his tumor which eroded the sacrum, 
from colloid carcinoma, but I have seen all these features, except the broad 
attachment to the sacrum, in colloid carcinoma ot the rectum. In the 
pharyngeal region atypical chondromas are rather common and must be 
considered a more probable occurrence than chordoma. 


Circumscribed overgrowth of bone occurs under sucli a wide variety of 
conditions and the distinctions between inflammatory and neoplastic hyper- 
plasia of the tissue are so often obscure that it has never been possible to 
exactly define the limits of osteoma. True progressive neoplasms which 
adhere to Virchow's criteria and in which bone is the esse^itial and not the 
secondary or accidental product, are not common, but chronic processes 
which result in bone formation from trauma, inflammation and disturbances 
of nutrition are numerous. Volkmann concluded that all exostoses that 
arise from cartilage are true tumors, others not. Borst, who separates from 
the tumors all the bone-producing processes which are self-limiting or clearly 
associated with trauma and inflammation has little to say about true osteoma. 
Virchow and Ribbert largely agree in recognizing neoplastic qualities in many 
processes in which trauma and inflammation are originally concerned, and 
which, whether true tumors or not, illustrate the complex etiology of 

Bone formation in necrotic tissue (brain, eye, kidney, aorta) ; in the floors 
of ulcers; in the course of syphilis, tuberculosis and chronic suppuration in 
and about bones; after fractures, in chronic arthritis; at the insertion of the 
overused tendons; in bursje; in the muscles of riders — are some of the proc- 
esses in which neoplastic qualities are least prominent. Yet in all of them 
the size of the resulting bony mass, the long duration of the process, and its 
eventual independence of the original exciting factor occasionally reveal 
definite neoplastic characters. Hence in discussions of osteoma it is custo- 
mary to include all forms of overgrowth of bone, thus securing an effective 
comparison of true and partial bony neoplasms. 

Histological study fails as a rule to distinguish simple hyperplastic 
growth of bone from true osteomas. When originating in bone both processes 
show a participation of some or many osteoblasts which surround the edges 
of the newgrowth and add to its substance from one or many sides. Both 
processes yield dense lamellated bone with few Haversian canals or spongy 
bone with many vessels and abundant marrow spaces and cells. Bone 
forms in connective tissue without the appearance of many osteoblasts, 
the fixed cells being passively incorporated in osseous matrix, but this 
character does not always distinguish self -limiting from extensively pro- 
gressive osteosis. 

The gross and chnical features seem to form the best criteria by which to 
separate osteoma from simple hyperostosis. 

Spontaneous or traumatic but non-inflammatory origin, progressive 
course, circumscribed form, active participation of osteoblasts, and derivation 
from cartilage, are features which are more prominent in true osteoma, 
while an inflammatory origin, self-limitation, multiphcity, diffuse form, 
reduced numbers or absence of osteoblasts, and origin from connective or 
other soft tissues, belong to the less definite or spurious bony neoplasms. 

Terminology. — Osteoma is the term apphed to tumors composed of bone. 
Strictly speaking, it should be hmited to true neoplasms. 



Exostoses are circumscribed masses which project above the bony surface. 

Enostoses He within compact or cancellous bone. 

Hyperostoses are more diffuse enlargements of bone. OsteopTiyte is the 
general term applied to inflammatory periosteal bony deposits. 

The new tissue may be ivory-like with soUd lamellae with few or no 
Haversian canals (extososis eburnea) ; or spongy with cancellous tissue (ex- 
tososis spongiosa); or it may contain wide marrow cavities or spaces (ex- 
tososis meduUaris) . The growth may affect bony structures or other tissues, 
and is thus hyperplastic or heteroplastic. 

The osteophyte is an inflammatory cortical or supracortical deposit of 
bone from the periosteum. It appears as a flat, often extensive plate 
or more circumscribed mass deposited on preexisting bone from which it 
may usually be detached. Microscopically it consists of many small islands 
and thin trabeculae of bone lying in the cellular and vascular connective 
tissue of the thickened periosteum and continuous with the preformed bone 
from which it springs. Many stages of the process of ossification are visible 
in early cases. 

Any area of periosteum which is affected by chronic inflammation may 
produce osteophytes. They are most frequently observed on the shafts of 
long bones; in the neighborhood of chronic sinuses or ulcers; as a prominent 
feature in the course of chronic arthritis; on the maxillae about inflamed teeth; 
and beneath the dura of the frontal or occipital bones in gestation. 

The course of inflammatory periosteal bony deposits varies with their 
etiology. Most cases reach a standstill as the inflammation subsides, leaving 
deformities and interference with function. Ankylosis of joints may result. 
In not a few cases osteophytic growth seems to have led to progressive over- 
growth and the production of extensive tumor-like processes. In the case 
of Forcade's son, cited by Virchow, suppuration in the lachrymal region led 
to progressive enlargement of many bones of the face and skull continuing 
for 33 years, and resulting in extensive leontiasis ossea. 

Hyperostosis. — Diffuse hypertrophy of large portions of bones, or whole 
bones, or several bones, occurs under many conditions. The most famiUar 
form of hyperostosis results from osteomyelitis, periostitis, and arthritis. 
While the hypertrophies of this class may be extensive they are readily 
separated from neoplastic growths. In acromegaly there is diffuse over- 
growth of bones of skull and extremities resulting from disturbance of func- 
tion of the hypophysis (leontiasis ossea) . The old and frequently cited case 
of Sancerotte of general leontiasis ossea is unique but of uncertain nature. 
The bones of the face and skull are specially prone to diffuse enlargement 
from local factors, trauma, inflammation, and rickets, and not all cases of 
craniosclerosis can be referred to acromegaly. Virchow collected many 
cases of diffuse enlargements of temporal, sphenoid, maxillary, and malar 
bones, and head of femur, apparently resulting from varied causes, not dis- 
tinctly inflammatory. He considered these overgrowths of bone as com- 
parable to elephantiasis of the skin and lipomatosis. 

Exostoses. — The circumscribed tumor-hke masses projecting from the 
surface of bones and intimately connected with the shaft, present a varied 
form, course, and etiology. They are flat or tuberous, nodular, globular, or 
pointed or ragged growths, which appear in many striations, at epiphyseal 
junctions or on the shafts of long bones, and at the insertions of tendons and 
fascia. They are flat, smooth and ivory-like as is usually the case with 
exostoses of the skull (ex. clavata) ; or spongy with cancellous spaces com- 
municating with those of the shaft. The porous exostoses may later become 
soUd. They are usually multiple, often symmetrical, and some notable cases 


are recorded in which there was a prominent hereditary tendency transmitted 
through the males (Heymann, Reinecke). In a case of hereditary multiple 
exostoses E. Miiller found very numerous islands of cartilage in the perios- 
teum of long and flat bones. 

In advanced forms the general development of the bones has been retarded 
and even dwarfism has occurred (Virchow). As a rule exostoses occur in 
young subjects before puberty during the period of active growth of the 
bones (exostoses de croissance). 

Exostoses usually arise from overgrowth of subperiosteal bone and these 
must be attributed to an irritation afl'ecting the periosteum. Yet the entire 
thickness of the shaft may be involved and new cancellous spaces of the 
exostosis communicate with the old. Or the newgrowth may lie within the 
bone shaft or marrow cavity (enostosis). In other cases the process is 
entirely superficial and pejtiosteal and some of these may lie within the perios- 
teum and apart from the shaft (parosteal exostosis). Trauma, with or with- 
out misplacement of a fragment of periosteum, or fracture, is often the ex- 
citing cause of such growths. About the ends of long bones portions of 
cartilage may be found in exostoses (ex. cartilaginea). An origin from a 
inisplaced island of cartilage is probable, and may be referred to trauma, 
rickets, or preexisting ecchondrosis. Microscopically exostoses usually begin 
with prolifejEation of the inner vascular layer of the periosteum and the forma- 
tion of a network of osteoid tissue. This matrix may at once become eburn- 
ated, or vascular channels may persist and the tissue remain spongy. The 
underlying bone becomes sclerosed and fused with the newgrowth, or it 
becomes rarefied and the adjacent marrow spaces communicate with those 
of the exostosis. According to Virchow the former process may be called 
histioid, the latter organoid. 

Clinical Groups of Exostoses. — The largest class of exostoses are those 
which form at the ends of long bones and are connected with disturbances of 
growth. They are multiple, tend to grow less rapidly as the subject reaches 
puberty and rarely they may spontaneously disappear (Hartmann). The 
long bones, the vertebrae, pelvis, scapula and the skull are the chief sites 
affected. Partly cartilaginous exostoses occur about joints and may possess 
a synovial covering which is derived from and may communicate with the 
joint (exostoses bursa ta) . Some of these are probably originally intra-articular 
and gradually become separated from the joint cavit)^ (Bergmann, Volkmann) . 

Multiple exostoses occur on the bodies of the vertebree and involving the 
intervertebral discs. In the cervical region they may produce palpable 

Traumatic exostoses constitute a well-defined class. They occur at the 
seat of fractures when excessive callus forms and after single or repeated 
trauma of the periosteum or bone and along the alveolar borders after ex- 
traction of teeth. Virchow cites two cases where broken splinters of bone 
hypertrophied, forming exostoses of moderate size. The great toe is the seat 
of a subungual exostosis which may occasionally develop into a true osteoma 
or osteochondroma. Continuous mild trauma is also responsible for the 
discontinuous exostoses which form in the deltoid in soldiers where the gun 
rests on the shoulder, and in the adductor muscles of riders. 

In nearly all these situations occur larger growths with more distinct 
neoplastic characters suggesting that the small exostoses may occasionally 
give origin to true osteomas. 

In the trachea Ribbert and Mischaikoff have fully described the multiple, 
round, flat or ragged nodules and plates of bone which grow in the submucosa, 
and have found that most of them, like the tracheal ecchondroses, are con- 



nected with the perichondrium of the tracheal rings. In a very advanced 
case described by Heyman, with extensive ossification in the larynx, there 
were many bony masses with marrow spaces throughout the trachea. 


Osteoma of Facial Bones and Antra.— The bones of the face are the seat 
of a variety of bony overgrowths which are difficult to classify. 

In the maxUlce osteosclerosis and hyperplastic bone formation occurs m 
connection with dislocated and inflamed teeth producing exostoses or enos- 
toses, or hyperplastic masses some of which project externally or into the 
antrum of Highmore. Many of these must be regarded as inflammatory, 
while others are true osteomas. 

, -^i,'.* >Jf ^J » ^ ■-" '/ 5 ?' / ,' '/ -^ I..- \ -,. . »*•» 

t ■' ' ■■•■■ 


*.; -'». 

Fig, 45- — Growing osteoma of antrum. 

In the orbit bony growths arise from any of the bones of this cavity, usually 
from the upper and inner segment. 

They not infrequently appear as congenital tumors in young girls. 
Virchow concluded that they arose as enostoses of the orbital wall and grew 
either into the orbit or nares or into the cranial cavity. He described four 
cases, in some of which cysts were found lined by ciliated epithelium. Many 
bony tumors have been described as arising from the ethmoid, frontal, and 
superior maxillary bones, projecting into the cranium, nares, frontal sinus, 
or antrum of Highmore. 

LaGrange has collected 148 cases. Their origin appeared to be from the 
periosteum or, as Lesser states, from embryonal cartilaginous portions of the 
ethmoid. Projecting into the sinuses they present a covering of inflamed 



and often cystic mucosa (Panas). The chief site is the upper inner wall 
of the orbit. They may penetrate the skull, but seldom cause grave cerebral 
symptoms, while only two such cases were fatal. Knapp has reported two 
enucleations, in one of which the growth was broadly attached over the upper 
wall of the orbit. 

Osteoma of the frontal sinus is believed to arise from fragments of carti- 
lage connected with the ethmoid. It produces a characteristic swelling over 
the orbit and progresses steadily until extirpated. Its structure is usually 
solid and lamellated. Most of the osteomas of the nares and antrum of 
Highmore also arise from the ethmoid (Bornhaupt). Tillmanns collected 
several cases in which these tumors were found loose in the nasal cavity or 
antrum (dead osteoma). 

Fig. 46. — Osteoma of orbit. (Afler^Knapp.) 

Secondary Osteoma. — Osteoma occurs as an element in many other 
tumors where it appears as a result of secondary ossification of the connective 
tissue. Here it may signify a terminal product in a series of transformations 
of the tissues of fibroma, lipoma, sarcoma (Borgt). In periosteal fibroma 
bone is a natural end product of the growth (fibro-osteoma). 

In certain cases it represents an integral part of a mixed tumor. In the 
common breast tumors of dogs, bone formation may be very active and ex- 
tensive, producing complex osteo-chondro-carcino-sarcomas. In the ovary 
osteoma occurs in several forms (Kroemer, Lit.) . True bone with periosteum 
and marrow spaces is observed in connection with dermoids and teratomas. 
Ovarian stones composed of a thick shell of bone about a central cavity arise 
from ossification of cysts, some of which are from corpora lutea. They are 
usually preceded by active inflammation. Metaplastic bone formation 
occurs in ovarian fibromas and chondromas, either in small foci or throughout 
the tumor. 


Osteoma in the Nervous System. — In a considerable portion of cachectic 
subjects, especially of the insane and epileptic, there are multiple nodules of 
bone in the cerebral or spinal arachnoid. Zanda finds that these nodules 
form about small vessels through degeneration, exfoliation and sclerosis of 
endothelial cells and connective tissue, which form hyahne, then osteoid and 
finally bone tissue. From these nodules or over large areas, especially in the 
falx cerebri and cerebral and spinal dura, larger masses of bone may form 
with cellular and fatty marrow spaces, yielding bony plates which are smooth 
and incorporated in the dura externally, and rough internally where growth 
progresses (Zanda, Ziegler, Borst). 

Larger tumors compressing the brain or cord and lying within the brain 
tissue or arising from the membranes have been described by several authors 
(Virchow, Meschede, Ebstein, Bidder, DeVecchi). 

The intracerebral growths were referred by Virchow to osteogenesis by 
the gha-tissue. Some are associated with encephalitis with softening. 
Others doubtless originate from misplaced islands of the bone-forming matrix 
of the dura. 

Heteroplastic Bone Formation. — In heteroplastic bone formation the 
process begins either in a cartilaginous or a fibrous matrix. The cartilage 
becomes more vascular and ossification occurs about the small vessels through 
the activities of osteoblasts. In connective tissue the stroma becomes 
hyahne, calcification occurs under the influence of osteoblasts with the 
appearance of osteoid and finally osseous tissue. Here one must conclude 
that the process is metaplastic, the fibroblasts acquiring the function of 

Several factors may be regarded as tending to call forth osteoblastic 
properties in fibroblasts. 

1. Proximity to Bone. — Some authors have assumed that bone formation 
always results from osteoblasts that have wandered out from periosteum 
(Busch). In myositis ossificans this theory finds a certain support as the 
process begins in the periosteum. 

2. The presence of calcific deposits figures in many instances of ossification 
of necrotic tissue. 

3. An active productive inflammation with organization of dead tissue and 
blood-clot, is probably essential in the ossification of muscle-tissue after 
trauma (Busse, Berndt, Rapke, Lit.). 

4. A special predisposition to calcification and ossification, possiblj' con- 
nected with disturbance of calcium metabolism, must be assumed to exist 
in certain cases, notably in those of reticulated osteoma of the lung. 

In many cases heteroplastic bone formation occurs without the appear- 
ance of many osteoblasts, the fixed cells being passively incorporated in 
osteoid and osseous matrix. 

Clinical Forms. — The lung is subject to an interesting and rare type of 
bone deposit through the progressive ossification of inflammatory connective 
tissue. This condition was first described by Luschka and may be termed 
reticulated osteoma. 

The condition appears in three forms which are probably stages of the 
same process. It first produces multiple nodules or tuberosities in the con- 
nective tissue of one or, several lobes. 

In early cases in young subjects both lungs throughout presented about 
fifty pea- to bean-sized nodules with much new connective tissue (Wagner, 
Heschl). In later stages one or more lobes are the seat of extensive reticu- 
lated bony deposits radiating in stalactite form from the root of the lung 
(Forster, Bostrom, Picchini, Triboulet, Arnsperger). In a few cases one 


lobe is transformed into a nearly diffuse mass of bone while nodules appear in 
other lobes (Port, Cohn). 

Bone masses in the wall of a cavity with nodules in other portions of the 
lung were observed by Rulher, Browning and LeDiberder. 

Microscopically the process begins as an interstitial growth of connective 
tissue, in the septa, walls of bronchi and alveoH, and about the large vessels, 
all of which are inclosed and eventually occluded by ossifying connective 
tissue. The vessels may show endarteritis obliterans (Picchini). The 
islands of bone are surrounded by numerous osteoblasts and marrow spaces 
filled with fat and marrow-cells may appear. 

The majority of observers have concluded that the process, at least 
at its inception, is an ossifying interstitial pneumonia. In advanced stages 
it seems to acquire certain neoplastic characters, and Virchow regarded 
the diffuse form as a true osteoma. The reticulated osteoma is to be dis- 
tinguished from ossifying chondroma of the lung but in advanced cases 
this may be difficult to accomplish. 

Simple calcification may occur in the walls of cavities (Krauss, Nusser) 
and the lung is a frequent seat of metastatic osteosarcoma (Gross). 

In the penis of a subject 42 years old, Lenhossek has described the for- 
mation of bony and cartilaginous plates which probably resulted from the 
changes in inflammatory tissue in the fibrous septum and sheath. Konig 
refers to similar cases. 

In the skin multiple nodules of compact bone appear in the epidermis 
or derma, sometimes about cartilaginous matrix (Virchow). 

Coleman has described osseous plates in the sole of the foot. 

In the eye, the sclera and choroid, are occasionally seen bone plates which 
result from metaplastic changes in inflammatory connective tissues. 

Bone formation in necrotic tissue occurs in many situations, as in phthisis 
bulbi (Virchow), in the wall of the sclerotic aorta (Cohn, Klotz), in the walls 
of abscesses and in chronic pleurisy, and in advanced stages marrow spaces 
and cells may appear. After ligation of the main vessels of the kidney 
there may be extensive formation of bone and cellular marrow in the necrotic 
organ of the rabbit. 

Myositis Ossificans. — Spongy bone may form as a result of trauma in 
the muscles of working people and has been observed in the biceps, digastric, 
vastus externus, adductor magnus, and diaphragm (Virchow). In these 
cases the ossification is self-limiting but in not a few cases of traumatic 
origin there is a certain progressive tendency suggesting a relation to myositis 

The progressive ossification of one or many muscles, first fully described 
by Munchmayer and called progressive m3'ositis ossificans, arises spon- 
taneously, chiefly in young subjects and pursues a chronic intermittent 
course. It affects the shoulder, face, trunk and limbs, and may become 
quite extensive. In the early exacerbations there may be local congestion 
and pain. 

Once established it tends to progress until the affected muscle is com- 
pletely ossified and immobile. Yet Elliott observed complete resolution 
of the preliminary myositis. According to Ribbert the ossification emanates 
from the periosteum, spreading . along the fascia, tendons and the peri- 
mysium and endomysium, with atrophy of the muscle-fibers. 

Lexer finds first, increase of connective tissue, then chondrification and 
finally ossification, constituting three stages in the process. 

Since the disease occurs at an early age in subjects with deformities of 
fingers and toes, is associated with multiple exostoses, and progresses inde- 


pendently of any originating trauma or other known factor, Ribbert regards 
the process as blastomatous. Since the intermuscular connective tissue 
may be regarded as a part of the skeletal system and intimately connected 
with the periosteum (Koester), it may be supposed that in these cases the 
former tissue retains more of the periosteal character than is normal, thus 
constituting a tissue predisposition toward bone formation (Mays). The 
occurrence of very similar lesions in tabes and syringomyelia suggests that 
a spinal trophic influence may be concerned in some cases (Kaufmann). 


^ Myoma, a tumor of muscle-tissue, occurs in two types, leiomyoma con- 
taining smooth muscle-tissue, and rhabdomyoma composed of striated musde- 
tissue. In both cases supporting connective tissue and blood-vessels usually 
accompany the muscle-cells and the structure is organoid. Yet pure 
leiomyoma occurs in the gastro-intestinal wall and the early stages of many 
myomas contain Uttle or no connective tissue. According" to the rule with 

^ j^ ■* «^ /» / «*" V. I f® «• '*'^' h r 

Fig. 47- — Structure of simple benign myoma uteri. 

benign turnors, myomas are easily recognizable as such, and metaplasia 
and infiamrnatory factors play a very subordinate role in their history. 


Leiomyoma occurs in the form of single or more often multiple, miliary 
•or voluminous, firm and opaque, or cystic tumors chiefly in preexisting 
smooth muscle-tissues. Some contain much connective tissue and are 



quite hard, leiomyoma durum; others are succulent and edematous, 
leiomyoma molle; rarely dilated blood-vessels are prominent, leiomyoma 
cavernosum; and occasionally cysts of dilated lymph- vessels, glandular 
structures, or softened edematous areas may form. Extreme congestion 
may render them dark and mottled, and after strangulation of vessels 
they become black. Fatty degeneration may produce yellowish soft areas, 
and calcification may transform entire tumors into stony masses. 

On section leiomyoma presents characteristic striated or convoluted 
markings from intertwining bands of muscle-cells. The tumor is more 

Fig. 48. — Multiple myomas of uterus, one cystic. 

opaque than normal muscle-tissue and is usually sharply circumscribed, 
often shelling readily out of its bed. The curved markings .often center 
about blood-vessels from which the growth appears to originate. In injected 
specimens Ribbert finds the blood-vessels small amd scanty as compared 
with surrounding tissue, and the nutrition is maintained largely by lymph- 
vessels. Hence myomas grow slowly and are very subject to standstill 
and regression. 

The structure of leiomyoma presents a system of intertwining bundles 
of muscle-cells supported by septa and fibrils of connective tissue. In 

MYOMA 201 

some cases the muscle bundles are definitely related to small blood-vessels 
or glandular structures but more often no such relation is apparent. The 
edges of the tumor are usually sharply defined but may grade insensibly into 
the surrounding tissues. The tumor-cells are thicker and shorter than 
normal smooth muscle-cells, the cytoplasm grows less acidophile and the 
nuclei are larger and richer in chromatin. In actively growing tumors 
there is increase in the neoplastic characters of the cells, especially in their 
size, and in malignant leiomyomas very large spindle-cells appear, giant-cells 
form of enormous dimensions, and the connective-tissue stroma disappears. 

The diagnosis between leiomyoma and spindle-cell sarcoma is often 
difficult and may be based on the clinical history, the acidophile character 
of muscle cytoplasm, the long oval nuclei, the arrangement of the cells 
and the presence of many adult collagenous fibrils running between the cells. 
Both cells and nuclei are more uniform in size and exhibit more evenly rounded 
ends than fibroblasts. Special stains facihtate the diagnosis. The ap- 
pearance of cross-sections of muscle bundles with nuclei embedded in 
acidophile fibers, alternating with longitudinal sections, is usually quite 

The connective-tissue stroma is acellular, often edematous or infiltrated 
with round-cells. Two types of blood-vessels are found in most myomas. 
In one the walls are normal, while the others participate in the neoplastic 
process, the walls are composed chiefly of hypertrophic muscle-cells and 
the adventitia is deficient or absent. 

Regressive changes are frequently present. Fatty degeneration affects 
individual cells in moderate degree, and extreme fatty and hydropic degenera- 
tion or necrosis with edema produce softened areas. Hyaline degeneration 
may overtake small areas or an entire tumor. Calcification follows fatty' 
degeneration or hyaline change and may become universal. Inflammatory 
changes in myomas are usually slight. Round-cell infiltration of the stroma 
is associated with edema, and about the included glandular structures of 
some myomas there is adenoid tissue, or lymph-follicles may develop. Mast 
cells may be very numerous (Reich, Gottschalk). Occasionally myomas 
become infected and infiltrated with polynuclear leukocytes, and cystic 
myomas may be found distended with pus. 

The course of leiomyoma is slowly progressive over a period of years 
and the symptoms are mainly of mechanical origin, or referable to associated 
conditions. There is a natural tendency for many leiomyomas to reach a 
standstill after a long period of growth or to regress with atrophy of muscle- 
fibers, replacement fibrosis, hyaline degeneration and calcification. In 
the uterus this tendency reveals itself especially at the menopause, but it is 
not constant. On the contrary very rapid growth or sarcomatous trans- 
formation may begin at the menopause. Gibson observed the appearance 
of a uterine myoma and growth to the size of a child's head within seven 
months after removal of both tubes and ovaries. Regressive changes may 
be hastened by pedunculation, by hyaHne changes in stroma, by closure 
of vessels, by separation from attachments, and by inflammatory processes. 
Acceleration of growth may result from abnormal adhesions. 

Myoma Malignum. — In a group of cases now rather numerous, 
leiomyoma has proved malignant, breaking its natural boundaries, and 
producing metastases in liver, 'lungs, kidney, peritoneum, and lymph-nodes, 
and thereby acquiring the designation myoma malignum. From the uterus, 
stomach, esophagus, and intestine this usually benign tumor has developed 
extensive local and general secondary growths and proved fatal. This event 
has occurred in the later periods of the growth or after extirpation of tumors 



which had long remained benign so that it would appear that a malignant 
change had taken place in the character of the tumor. The symptoms 
observed are a rapid increase in the size of a long quiescent growth, signs 
of secondary growths, ascites in the uterine cases, and cachexia. The main 
tumor has usually been adherent, with cavernous, softened, or cystic areas. 
In one of my cases the softening spread from a crater-Hke ulcer in the adherent 
endometrium. In Schlagenhaufer's case it was limited to an angiomatous 

In most of the cases the structure of the soft pultaceous areas and of 
the metastatic growths which probably emanate from these areas, differed 
from that of ordinary leiomyoma. The cells were larger, the nuclei more 
massive and hyperchromatic, mitoses were present, the fusiform cells became 

Fig. 49. — Miliary myoma in a uterine vein. The tumor is benign. 

shorter or rounded, giant-cells formed, intercellular connective tissue was 
scanty or absent, and the walls of vessels were defective (Krische, 
Langerhans, Williams, Beeston, Moser, Basso). In Hansemann's malignant 
myoma of the stomach the cells were uniform but short or rounded and the 
nuclei very rich in chromatin. In Eising's esophageal myoma islands of 
adenocarcinoma were present and the muscle-cells much increased in size. 
In one of Evelt's cases the sarcoma was apparently melanotic. In three 
uterine cases I found the structural changes very marked. In one (3461) 
there was extensive mucoid degeneration. In another (2287) the spindle- 
cells and giant-cells reached enormous dimensions. In all the derivation 
from muscle-cells was clear. So far as I have been able to learn no case has 
been fully studied in which definite variation from the usual structure of 

MYOMA 203 

leiomyoma were wanting, although in several instances these variations have 
not been ver}- pronounced. 

_ In one case of large uterine myoma I found the peritoneum of the uterus 
thickly beset with miliary myomas identical in structure with ordinary 
myoma, and in a case of recurrent multiple nodular myomas of the broad liga- 
ment the cells varied little from the usual type although they were shorter 
and looser and the tissue was very vascular. In neither case were there dis- 
tant metastases. Accordingly it may be said that the structure of malig- 
nant myoma varies from that of the benign type. 

Ribbert has protested against the term myosarcoma, employed by many 
for the malignant tumors, on the ground that sarcoma implies a tumor of 
connective tissues. Yet in a morphological sense this term seems admissible. 
Objection has also been made to the interpretation of the above changes as 
genuine sarcomatous degeneration of myomas. There has been a tendency 
on the part of gynecologists to multiply the cases of this sort, and to search for 
suspicious foci over wide areas of perfectly benign tumors. Some estimates 
place the frequency of the change at lo per cent, of all uterine myomas. 
From the reports of Winter, Henkel, Hofmeier, and Flautau, in 1880 cases 
of fibroids 48 or 2.5 per cent, were found to show sarcomatous changes. The 
manner in which these statistics were obtained is perhaps fairly indicated in 
the study of Winter. In his first series of 500 cases he examined only sus- 
picious-looking areas and found 16 (3.2 per cent.) to contain sarcomatous foci. 
In 253 later cases he examined all parts of the tumor and succeeded in rais- 
ing the number of sarcomas to 1 1 (4.3 per cent.) . There were sometimes pres- 
ent general signs of malignancy, as rapid growth, polypoid masses on the 
main tumor, and changes in consistence, but local metastases occurred only 
once and cachexia resulted chiefly from hemorrhage. In the same manner 
CuUen has reported 17 cases of sarcomatous degeneration of uterine myomas. 

These observations show that ordinary myomas vary in structure in 
different portions and probably at different periods, and they seem to justify 
the suspicion with which the gynecologist regards the entire group. But it 
must be considered that these suspicious changes may not always be pro- 
gressive but may signify merely a temporary or local acceleration of growth 
which may subside and even regress. They do not seem to justify their 
designation as sarcomatous transformations, for which much more exten- 
sively altered areas or even general metastases might well be demanded. 
Sarcomatous tendencies and precancerous changes do not constitute real 
sarcoma or cancer. I have encountered three malignant uterine myomas 
with general metastases and two with local [recurrence in 20 years, and 
Winter found no case among 753, so that malignant degeneration of myoma 
must be rare. 

It has been claimed that certain myosarcomas arise through a malignant 
transformation of the connective tisue of leiomyoma. (Ricker, v. Franque 
and Hauser). In Ricker's case the change occurred in numerous foci 
sharply separated from the muscle-tissue which was either infiltrated with 
the sarcoma-cells or underwent atrophy. Yet the descriptions of these 
authors are not entirely convincing that they were not dealing with the 
ordinary type of malignant change in muscle-cells. 

Clinical Types of Leiomyoma.- — Uterine Fibromyoma. — The uterus is 
the commonest seat of leiomyoma. The tumors are usually multiple, in 
fact in several cases the entire uterus has been found to consist of a congeries 
of discrete myomatoid masses (CruveiUiier, Williams, Stone). The lo- 
cation is widely distributed, the posterior wall, anterior wall, sides, and cer- 
vix appearing in the order named. From 5-8 per cent, of these tumors are 



cervical (Fehling). Extra-uterine tumors of the same type and significance 
occur also in the broad hgaments, vagina, tubal junction, in the pelvic fas- 
ciae, and in the round hgament as far down as the groin or labium majus 
(CuUen, AschoS). The position of the uterine tumor may be subserous, 
interstitial, or submucous. J^oth subserous and submucous tumors tend to 
become pedunculated or even separated from the uterus, being rarely dis- 
charged per vaginam (Simpson), or lying free in the peritoneum (Turner), 
or becoming attached to surrounding organs. Submucous tumors may gradu- 
ally separate from their original site and become implanted on the opposed 
mucous surface (Leyden, Kustner). Diffuse myomatosis may cause thick- 
ening of a small focus or of a large area, or of almost the whole organ. Mil- 

FlG. 50. — Structure of a multiple nodular recurrent myoma of broad ligament. 

iary myomas appear in the peritoneum during the course of larger tumors, 
while Hunter found at autopsy a single tumor weighing 140 pounds. 

The form of uterine myoma is usually a rounded, sharply circumscribed 
solid and elastic mass which projects prominently from the cut surface. 
Many very early myomas and many adenomyomas are without definite 
demarcation and the diffuse forms remain so, but with increasing age most 
tumors show more and more isolation and some become encapsulated. 
Cysts, multiple and superficial or large and central, form in many myomas 
from dilated gland alveoH, or from chronic edema and mucinous softening. 
Their contents are serous or mucinous fluid with cholesterin, or after infec- 
tion they may become distended with pus. Some of the cystic myomas reach 
a very large size (Martin), that removed by Severanu of Bucharest being one 
of the largest tumors ever observed. They may contain pus or chocolate- 
brown or bloody fluid. The uterus which is the seat of myoma, especially 

MYOMA 205 

of the submucous or interstitial varieties, enlarges with the growth of the 
tumor and becomes extensively deformed. There is hypertrophy and h)qDer- 
plasia of the muscle-cells as a result of contractions caused by irritation from 
the tumor (Bertelsman), but Tillaux saw the uterus greatly enlarged with a 
small tumor of the lower segment. The vessels about the tumor are enlarged 
and the whole organ is hyperemic. Endometritis, interstitial or glandular, 
commonly results, especially with submucous varieties, and mucous and bloody 
discharge becomes a prominent symptom. Inflammatory hyperplasia may 

Fig. 51. — Myoma of uterus, confined under pressure with edema, small cysts, and polypoid 

growths into cysts. 

be hmited to the mucosa overlying the tumor (Wyder), or it may extend also 
to the tubes (Fabricius) and ovaries (Bulius). Profuse hemorrhage between 
or prolonging the menstrual periods marks the course of many forms of my- 
omas, and sometimes proves fatal. According to the position of the tumors 
the uterine cavity becomes elongated, irregular, or occluded. The stretch- 
ing of the myometrium may produce peculiar rhombic spaces in the stroma 
and the stretching has even led to rupture of fundus from the cervix (Len- 
nander, Hedren). Mural myomas may cause extensive torsion of the uterus, 
and submucous tumors may lead to prolapse or inversion (Kuster, Kotschau). 



The gross section reveals the usual convoluted markings of myoma, or a 
congeries of multiple tumors. Islands of adenoid tissue and the sections 
of glands may be revealed by the hand glass. Multiple minute cysts appear 
when the alveoh are moderately dilated. The tumor-vessels are usually 
scanty while the surrounding tissue supports numerous dilated veins, but 
small areas (CuUen), or the entire tumor (Virchow, Martin), maybe telan- 
giectatic. Dilated lymph-vessels were prominent in two myosarcomas de- 
scribed by Menge. Submucous myomas are often traversed by glandular 
tracts, the openings of which may be seen on the mucous surface. 


Fig. 52. — Structure of a locally infiltrating myoma uteri. 

Extensive degenerative changes frequently alter the appearance of uterine 
myomas. Marked edema produces softening and cystic areas. Chronic 
edema leads to extensive mucinous degeneration with increase in the size of 
the tumor and formation of cysts. With increasing age there is a tendency 
toward atrophy and fibrosis which may greatly reduce the size of large tu- 
mors and render others very hard. Fatty degeneration of muscle-cells is 
associated with their atrophy and probably precedes calcification. Fatty 
changes may become very active during gestation and Martin saw a large 
myoma almost wholly transformed into fatty semifluid material during ges- 
tation arid the puerperium. Large tumors undergo fatty degeneration and 
may be reduced to small nodules or entirely disappear (Duncan, Hewitt, 
Gusserow). Hyaline degeneration may affect the blood-vessels, stroma, or 
muscle-tissue, in small foci or in considerable areas. According to Ribbert 

MYOMA 207 

this change may precede calcification, the salts being deposited in the hyaline 
material. Stratz saw amyloid degeneration about the vessels of a polypoid 
myoma. Calcification overtakes fatty and atrophying myomas and com- 
pletely transforms them into stony masses. Uterine stones of large dimen- 
sions are thus found in the cavity, wall, or on the peritoneal surface of the 
uterus (Henocque), and they have been discharged through the vagina or 
intestine, or into the bladder (Payr). Everett found the composition to be 
calcium carbonate 49 per cent., phosphate 29 per cent., sulphate 13 per cent., 
with traces of lithates and organic material. Turner and Wylie found a com- 
plete separation from the uterus of calcified subperitoneal myomas. Throm- 
bosis of vessels is not an uncommon event, especially with pedunculated myo- 
mas, and leads to hemorrhagic infarction, necrosis, softening and even to 
excavation and discharge of large portions of submucous tumors. 

Sarcomatous changes, previously considered, occur in a small proportion 
of uterine myomas. Carcinoma is sometimes observed with uterine myoma 
and may result from several conditions. In adenomyoma, carcinoma may 
develop from the epitheHal elements in the tumor. A submucous myoma 
may be complicated by carcinoma of the overlying and long inflamed mu- 
cosa, and the carcinoma may penetrate the myoma. Or the carcinoma may 
arise in a distant portion of the mucosa. Carcinoma of other organs has 
been known to produce metastases in uterine myomas (Liebmann, Schafer, 

The microscopical structure of uterine myomas reveals two forms of this 
tumor, fibromyoma and adenomyoma. The great majority of the tumors 
are of the former class and present the usual characteristic structure of fi- 
bromyoma. At their inception these tumors are nearly pure myomas and 
some remain so throughout most of their course but with increasing age there 
is a tendency toward increasing development of connective tissue which in 
atrophying tumors may exceed the muscular tissue. The connective tis- 
sue is distributed in septa supporting vessels and as finer fibrils ramifying 
between the muscle-cells. Gebhard observed a large, purely fibrous nodule 
in the center of a solid myoma. The connective tissue does not participate 
in the tumor process although some observers have claimed that it might 
give rise to sarcomatous changes in myomas. The tumor-cells are larger, 
thicker and more granular than normal uterine muscle-cells, and the nu- 
clei are larger, more rounded, and richer in chromatin. During gestation 
the cells of myomas increase in size. Their arrangement is in intertwining 
bundles running in all directions and without definite relation to any struc- 
tural unit. Certain large tumors are composed of a congeries of somewhat 
discrete nodules in the centers of which are blood-vessels, and in some early 
myomas the growth centers chiefly about blood-vessels from the walls of 
which the tumor seems to spring. In such cases the adventitia of the vessel- 
walls may be deficient (Rosger). It has been claimed that nerves have been 
demonstrated in uterine myoma (Hertz) but they are insensitive, and their 
existence has been denied. 

In rare cases islands of bone have been found in uterine myoma (Lebert, 
Freund); a nodule of cartilage has been encountered by Bennet; and in one 
malignant case of adenomyosarcoma Kaufmann found both bone and car- 
tilage. A fibrochondro-osteoma described by Kworostansky was possibly 
an atypical myoma suffering extensive metaplasia. 

Adenomyoma presents the usual structure of the muscle-tissue but it 
contains glandular alveoli which appear in various forms. They usually 
form small groups of acini lined by cubical or cyhndrical and sometimes cili- 
ated epithelium supported by lymphoid stroma. These acini are identical 



in appearance with fragments of uterine mucosa with which they are often 
directly connected. Or the cells are more embryonal in type, with relatively 
large nuclei, with or without ciha, and the lymphoid stroma may be present 
or absent. Most observers have been unable to establish any constant dis- 
tinctions between the epithehal structures in adenomyomas in different lo- 
cations but the development of the glands varies widely from a few scanty 
acini to large cystic tumors composed largely of epithelial lined cysts. Pap- 
illary projections may grow into the cysts and CuUen saw adenocarcinoma 
developing from the papillae. During menstruation there is congestion 
and hemorrhage into the glands (CuUen), and Amos has described decidual 
changes during gestation. 

The occurrence of adenomyoma is relatively infrequent, but its distri- 
bution is widespread. * They may be submucous, interstitial, or subperito- 
neal, and they have been found in the broad ligament, round Ugament, 
and groin (CuUen), cervix, vaginal wall. Fallopian tube and rectum. 

Etiology. — Uterine myoma is the most frequent of all tumors, its occur- 
rence being estimated at 50 per cent, of all women over 50 years of age (Klob), 

Fig. 53. — Structure of adenomyoma of broad ligament. 

and at 20 per cent, for those over 35 years. The incidence reaches its acme, 
38.8 per cent., between the ages of 30 and 40 years (Gusserow). The 
negress is especially' susceptible. Before puberty it is extremely rare but 
Tillaux reported in a girl of 19 a cervical myoma which had given 
symptoms for six years. Leopold's statement that the rudiments of myomas 
may be found in the uterus in children has not been verified. An hereditary 
influence seems to have existed in certain families in which several members 
suffered from myoma at an early age (Veit, Gusserow). 

Sterility has long been held to be a factor in the etiology of myoma and 
Veit concludes that abnormal excitation and congestion of the organ without 
conception may excite tumor growth. It is highly probable that myoma 
favors sterility from the mechanical and inflammatory effects of the tumor. 
Both sterility and myoma are favored by the infantile type of uterus and by 
misplacements and other defects in the organ which go with each of these 
conditions. Virchow held that myoma is caused by local irritation. Many 
authors have pointed out sources of such irritation and shown that these 
act as secondary and exciting factors, which are not constant or essential. 



The essential factor in the etiology of myoma is an embryogenic dis- 
turbance in the structure of the uterus. The remarkable degree of isolation 
of many myomas, their widespread occurrence apart from the uterine body, 
and the presence in many cases of heterotopic inclusions, epithelial, carti- 
laginous, osseous, fatty, and rhabdomyomatous, clearly point to an embry- 
onal origin. Moreover, as Williams has emphasized, uterine myomas are 

Fig. 54. — Diagrammatic representation of development of genito-urinary tract. {After 


The Miillerian duct and derivatives are jet black. The Wolffian body and derivatives 
include all other structures except the sex glands, bladder, prostate and rectum. 

I, Indiiferent type. 2, IndiflEerent type, later stage, the Wolffian and Miillerian ducts and 
primitive ureter now opening into the urogenital sinus. 3, Male type, lower ends of Miil- 
lerian ducts fused to form the sinus pocularis. 4, Female type. 

often associated with a large number of abnormalities in the genito-urinary 

The nature of the embryogenic disturbance varies with different tumors. 

(a) It is generally agreed that the common pure fibromyoma results from 


a disturbance in the formation of the tubes, uterus, and vagina from the Miil- 
lerian ducts, which spht off from the Wolffian ducts at an early period and 
fuse to form the genital canal. 

The relation of certain early myomas to the blood-vessels of the uterus has 
long impressed many observers and suggested that uterine myomas arise 
from disturbances in the growth of the blood-vessels from the walls of which 
the uterus and vagina originally receive their muscular tissue. Rosger, 
Kleinwachter, Sobotta, and others have traced the development of early 
myomas from the vessel- walls and concluded that the blood-vessels control 
the origin and growth of uterine myomas. This view has much evidence to 
commend it, although it is seldom possible to demonstrate such a relation 
in advanced stages of the tumors. Yet in many pure, in adenomatous and 
in telangiectatic myomas, the blood-vessels are composed of neoplastic 
muscular-tissue. In a case of multiple recurrent telangiectatic myoma of the 
broad liagment I found striking evidence of the origin of the tumor masses 
from blood-vessels. 

The cases in which the whole uterus is composed of a congeries of myomas, 
or where a myoma replaces the cervix (Landau) or causes uterus didelphys 
(Pick), strongly suggest an origin from misfitting vascular units which 
go to make up the uterine muscle. Likewise single isolated myomas seem to 
represent overgrowth of muscle from single vascular units which have not 
fitted the general texture of the myometrium. Into many such tumors the 
invasion of glands has been traced from the mucosa. There are many who 
believe that the great majority of myomas of all types originate in this way. 

(6) Regarding the origin of adenomyomas opinions are at variance and 
it is probable that no single mode of origin can apply to all these tumors. 

1. A Miillerian origin accounts most satisfactorily for the majority of 
uterine adenomyomas, especially for the submucous, interstitial, and diffuse 
varieties. Duplication of the duct, fetal budding of the epithelium, mis- 
placements of islands of fetal mucosa, postembryonal misplacement of 
gland tissue, and invasion of the myometrium by gland alveoli during in- 
flammation, have been observed or suggested as accounting for the origin of 
these tumors (Meyer, Pick, Hauser, Schroeder, Ruge). 

In many submucous or deeper adenomj'omas the epithelial structures 
have been traced to the uterine mucosa, on the surface of which they some- 
times open by patent canals (Ribbert, CuUen). The invasion of a tubal 
myoma by glands of the tubal mucosa has been traced in serial sections by v. 
Franque. The structure of the glands in most adenomyomas is an exact 
duplicate of the uterine glands and they seem to functionate as does the 
uterine mucosa. 

2. Mesonephric Adenomyoma. — The epithelial structures in certain adeno- 
myomas were interpreted by Babes and by Recklinghausen as remnants of 
the Wolffian body or mesonephros, and many have supported this view 
(Ricker, Schroeder, Pick, Schickele, Ernst). Recklinghausen believed the, 
glandular structures in certain cases presented the characteristics of the meso- 
nephros which forms a closed system consisting of a chief canal or ampulla into 
which many parallel smaller canals open. The smaller canals, starting in 
wide ampullae advance in the form of convoluted secreting tubules lined by^ 
cubical epithelium and empty into the chief canal while lined by high or .cili- 
ated epithelium. Vascular polypoid projections of cellular tissue along the 
canals in certain tumors he interpreted as pseudoglomeruli. Occasional 
deposits of pigment he compared with the pigment of Giralde's organ (rem-. 
nant of lower Wolffian tubules). 

Recklinghausen and Pick have attempted to divide adenomyomas into submucous 

MYOMA 211 

(central and ventral) and paroophoral (dorsal and peripheral), the former deriving its 
glands from the endometrium, the latter from the paroophoron, the remnant of the Wolffian 
body (lower tubules). Pick holds strongly to the dual origin of adenomyoma uteri, claim- 
ing that the tumors located in the broad ligament are of Wolffian origin and may be des- 
ignated as parovarian and paroophoral adenomyomas. Under the term adenomyoma 
psammopapillare he describes a remarkable multiple papillary tumor arising in the broad 
ligament which well illustrates Recklinghausen's voluminous paroophoral adenomyoma. 
As the chief histological distinction he states that lymphoid stroma usually but not always 
accompanies uterine glands in adenomyoma while the alveoli of Wolffian origin lie in 
immediate contact with muscle-tissue and lack the lymphoid stroma. Yet Cullen finds 
such lymphoid stroma in tumors of the round ligament. 

While it is probable that certain extra-uterine adenomyomas originate from the paro- 
varium or paroophoron the majority of observers have not been impressed with the meso- 
nephric characters of these tumors. Neither the location of the tumors nor the arrangement 
of the ducts nor the character of the epithelium seem to be sufficient evidence to prove the 
mesonephric origin, since very similar structures are seen in submucous tumors and in 
others for which a Miillerian origin is extremely probable (Meyer, v. Franque). Cohen 
saw the characteristic arrangement of tubules in a submucous tumor of which the epithe- 
lium was clearly derived from the endometrium, and he found the pigment in adenomyoma 
to be ferruginous while that of Giralde's organ is iron-free. Aschoff found no uniformity 
in the location of the so-called paroophoral type of adenomyoma, and Kossmann and 
Lockstadt found a general resemblance between uterine glandular tissue and the epithelial 
structures in all adenomyomas. Pick states that there are both adult and embryonal 
types of adenomyoma, and such a distinction appears to be not infrequent, but the dis- 
tribution of these types does not accord with the other features which might separate 
two varieties of adenomyoma. The predominance of muscular tissue is more readily 
explained by a Miillerian origin, for muscle-cells are scanty in the Wolffian structures, and 
Pick regards this element as quite secondary and unessential in the anlage of mesonephric 

3. Remnants of the Wolffian duct (Gartner s duct), seem to give origin to 
certain extra-uterine adenomyomas. It has been shown by Reckhnghausen, 
Ricker, Pick, and others, that the whole region of the uterovaginal canal 
may contain isolated epithelial groups which are probably derived from the 
Wolffian duct. Borst found heterotopic Wolffian alveoli in the iliac lymph- 
nodes. Klein traced a persistent Wolffian duct in two infants. The proxi- 
mal portion traversed the broad ligament beneath the Fallopian tube to 
the uterus, and terminated in the uterine muscle at the os internum. The 
distal portion passed along the side of the uterus to the cervix, curved back- 
ward over the fornix and passed downward in the wall of the vagina to the 
hymen. The canal was hned by cubical epithehum and supported by smooth 
muscle, and Klein concludes that this structure may give rise to various 
tumors along its course, as cysts, myoma, adenomyoma, and adenocarci- 
noma. Cervical and vaginal adenomyomas, especially, have been referred to 
this origin (Breus, Herpf, Pf annenstiel) . Yet, except for their unusual posi- 
tion, the same difficulties exist in proving this origin as for the mesonephric 
type, since the structure of these tumors does not differ essentially from that 
of the Miillerian type (Cullen). 

4. Besides the epithelial alveoh, subperitoneal myomas may contain 
acini lined by cubical endothelium derived from inclusions of the peritoneum 
(Aschoff, Meyer). 

Summarizing the evidence one may conclude that simple myoma uteri 
arises chiefly from a disturbance in the development of the tubes, uterus 
and vagina from Miiller's ducts, which often leads at the same time to gross 
deformities and infantile characters in these organs. 

Adenomyoma arises chiefly from the fetal or post-embryonal inclusion of 
MuUerian epithelium in the tumor process. Extra-uterine adenomyomas in 
the broad and round ligaments and cervix may also arise from the meso- 
nephric elements or from the Wolffian duct, but neither the position nor stru^c- 
ture of these tumors permits their positive identification in every instance. 


The growth of many myomas is controlled by their relation to blood-vessels 
from which they are derived. The chief exciting factor is interthittent hy- 
peremia connected with irregularities in the sexual functions. 

Myoma in the Alimentary Tract. — The esophagus was the seat of small 
multiple leiomyomas containing also striated muscle-fibers and ganglion- 
cells in a case described by Pickler. A larger single tumor was observed by 
Ilhg. Both were connected with the inner muscular coat. Eising's case 
was complicated with adenocarcinoma. 

In the stomach Miodowski observed a bulky myoma which led to 
severe hemorrhage. He found two similar cases in the literature. Hanse- 
mann has reported a malignant gastric myoma with metastases in liver and 
pancreas. In a large tumor studied by Cohen there -prere islands of pancreas. 
Laboulbene saw four small myomas in the same stomach. Many gastric 
sarcomas are probably of myogenic origin {q.v.). 

In the intestine leiomyoma is not uncommon. The tumors are single 
or multiple. Nazzari saw 40 small tumors in one case and 120 in another. 
Steiner found them either subperitoneal or submucous, the latter type caus- 
ing stenosis or intussusception, the former producing diverticula. In Mer- 
cer's case fatal hemorrhage occurred. Cohen found islands of pancreas in 
an intestinal myoma which probably belonged with the adenomyomas of 
Trappe. Both gastric and intestinal myomas are apt to be associated with 
uterine myoma, and they appear to originate from isolated segments of the 
musculature, connected with blood-vessels or epithelial structures. Accord- 
ing to Boetticher and Lode three types of intestinal myomas may be dis- 
tinguished: (i) Small multiple nodular or polypoid tumors arise from local 
proliferation in the muscularis; the mucosa is free; (2) broad thick tumor 
masses form in the muscle layers while the mucosa becomes adherent to the tu- 
mor ; (3) larger polypoid subserous myomas may project into the peritoneum. 
Dermatomyoma. — A wide variety of myomas is observed in the skin. 
They occur on the buttock and extremities, at embryonal fissures (Babes), 
in the scrotum and labia (m. dartiques) (Besnier, Forster, Challard). They 
may be associated with keloid (Babes), or 'lymphangioma (Axel, Key), 
neuroma (Czerny), and xanthoma (Chambard, Gouilloud). Politzer has 
shown that certain xanthomas result from fatty degeneration of smooth mus- 
cle-fibers. In some cases the blood-vessels are very numerous and neoplas- 
tic, producing angiomyomas (Hess). The tumors are usually small and 
multiple, firm, movable, cutaneous or subcutaneous, nodules or masses. 
They may be painful and tender. They usually occur in adults but Hess 
observed early cases and Marc describes a congenital lymphangiectatic my- 
oma. The origin has been traced to the arrectores pilorum (Judassohn, 
Walters) or to the cutaneous arterioles (Hess, Marc). Borst suggests that 
some of the dermatomyomas are really neurofibromas. Sobotka traced 
the process from simple hypertrophy of the arrectores up to tumor growth, 
followed by degeneration. 

In the urinary bladder and passages myomas constitute a considerable 
proportion of the tumors occurring in those organs. 

In the bladder they may reach large dimensions and their location is 
either subserous or submucous (Stein, Terrier, Hartmann). In an infant 
Kaufmann found a myoma of the trigonum containing cartilage. A large 
myocarcinoma is reported by Volkmann. In the ureter Buttner observed a 
large '''pure myoma. In the testis Rindfleisch described a myoma which 
contained nerve-fibers and ganglion-cells, and Becker reported one case of 
doubtful nature. 

In chronic prostatitis there is often extensive hypertrophy of smooth 

MYOMA 213 

muscle- tissue which may approach the neoplastic grade. True myomas of 
this organ have not been clearly separated from this inflammatory overgrowth. 
In the kidney small multiple myomas and myolipomas occur in capsule 
or cortex and appear to be derived from fragments of capsule tissue (Miil- 
ler, Larkin). Jacobsens has collected several cases which were associated 
with cerebral sclerosis. In the breast leiomyoma has been observed by Ab- 
ramow and Ribbert. 


Tumors of striated muscle are characterized by their rare occurrence, 
embryonal type, and common association with other constituents in mixed 
tumors. They occur in very wide distribution and, while the genito-urinary 
system furnishes the largest proportion, few muscular regions escape an 
occasional development of rhabdomyoma, and the homologous growths are 
fully equaled in number by heterotopic forms. They usually occur in early 
life, some are congenital, and a few examples have been first noted in advanced 
or old age (Wolfensberger, Fujinami). 

In form they appear as single or multiple, nodular or voluminous, flat or 
rounded, circumscribed or diffuse and even polypoid growths. They are 
usually soft and grayish on section, and markings produced by muscle bun- 
dles alternate with abundant connective-tissue stroma. Very cellular 
forms are more opaque, yellowish or reddish, soft and often diffuse. The 
mixed rhabdomyomas present the varied appearance of teratomas. A cys- 
tic rhabdomyoma of the elbow has been described by Billroth. 

The structure of rhabdomyoma presents chiefly a system of parallel 
bundles or intertwining strands of striped muscle-fibers, supported by adult 
or embryonal connective tissue. The fibers may encircle blood-vessels or 
other structures but are more often diffuse. The cells of rhabdomyoma sel- 
dom exhibit adult characters. They are usually thin and much elongated 
and rarely they are branching (Zenker, Billroth). Both long and cross 
striation may be pronounced or the cross striation may be present in a part 
of the cell only or be entirely missing in embryonal and spindle-shaped cells 
which still retain longitudinal fibrillation. Or the cefls may closely resemble 
smooth muscle-cells. In large vacuolated cells Ribbert describes concen- 
tric striation of the perinuclear cytoplasm. In a testicular tumor he ob- 
served long tubular fibers, the walls of which were of striated or hyahne mus- 
cle fibrils and the core of granular sarcoplasm with many nuclei in axial 
positions. The ends of the fibers may be swollen, rounded, and contain 
multiple nuclei. In the more embryonal types the cells are spindle-shaped, 
shorter, and even round, and they may lose their acidophile character and 
all resemblance to muscle-cells. Yet such marked anaplasia is seldom if 
ever universal, certain foci commonly retaining cells of definite myogenic 
character. Even the round-cells of lymph-node metastases may retain 
traces of striation (Wolfensberger, Eberth), or the metastases may consist of 
cells showing no features of muscle-cells (Benenati). In many myomas the 
cells contain an abundance of glycogen which is evidence of their embryonal 
character (Marchand). 

The nuclei follow a type of large vesicular chromatic bodies, single or 
multiple, which he in a central area of. granular cytoplasm. Or they may 
project from the cell border like nuclei of sarcolemma. A definite 
sarcolemma is wanting but traces of such a structure with its nuclei have 
been seen in rare cases (v. Franque, Marchand). The stroma is either 
loose adult connective, or embryonal, sarcomatous or myxosarcomatous, 
and in the teratomas bone, cartilage and various other tissues may be present. 


Blood-vessels are usually abundant and sometimes, as in a large polypoid 
vaginal tumor described by Kaschewarowa, they are overdeveloped. 

In parts of certain rhabdomyomas it is difficult to determine whether one 
has to deal with voluntary muscle-fibers which have lost their striation or 
with genuine smooth muscle-tissue. The undifferentiated cells may closely 
resemble smooth muscle. Moreover rhabdomyoma occurs especially where 
smooth muscle-tissue exists (kidney, testis). To many observers it has 
appeared that smooth muscle-tissue may give rise to striped muscle-fibers 
in tumors (Ebefth, Rindfleisch, Arnold, Marchand). In a myosarcoma 
uteri V. Franque saw islands of striated muscle together with smooth muscle 
and spindle-cell sarcoma and he concluded that the striped muscle arose by 
metaplasia of smooth muscle. Yet Girode and Nehrkorn both saw islands 
of striated muscle in the normal myometrium and the embryology of this 
organ provides the possibihty that portions of striped muscle-tissue may be 
included in its structure. Arnold, in supporting a suspicion that metaplasia 
of smooth may produce striated muscle, derived a rhabdomyoma testis from 
the cremaster muscle, but the tumor lay within the albuginea, while this 
muscle lies without the organ. The embryology of muscle- tissue, voluntary 
muscle being derived from mesothelial plates and involuntary muscle from 
the mesenchyme, does not favor the idea that one form may pass into the 
other. While many earlier observations suggested the change of smooth into 
striped muscle, most recent studies have not favored this view (Ribbert, 
Hauser, Kolisko). 

The course of rhabdomyoma is usually progressive and in the 
sarcomatous and teratoid forms it is rapid. The more adult types may 
become encapsulated or assume the polypoid forms, but the. embryonal 
tumors infiltrate surrounding tissue, multiplying muscle-cells being often 
preceded by a zone of proliferating connective tissue. 

Degenerative changes, glycogenic, hyaline or amyloid may occur without 
interfering with the growth. The malignant tumors of the kidney and 
testis reach large dimensions and produce metastases and cachexia, but 
the metastases arc usually derived from other elements of the tumor. 


Adenomyosarcoma of Kidney. — The kidney is the commonest seat of tumors 
containing straited muscle, but since none of them has yet appeared in pure 
form they will be described under tumors of the kidney. 

In the urinary bladder polypoid rhabdomyomas have been described 
by Cattani, Vincenzi, Stumpf, and Monckberg. They occurred in children 
or young adults, and formed movable growths about the urethral orifice 
or trigone. 

Rhabdomyoma uteri appears almost exclusively as an element in the poly- 
poid vaginal sarcoma of children and adults (Sarcoma botryoides, Pfan- 
nenstiel). This process affects the vagina in children and chiefly the cervix 
in adults (Gow, Pick). It may exist at birth (Demme) and first appears 
as a rather broad thickening of the submucosa which soon becomes polypoid. 
V. Franque reported a somewhat similar tumor' located on the fundus uteri 
in a multipara. The symptoms are hemorrhage, fetid discharge, and 
protrusion of a polypoid tumor from the vagina, with dysuria, pain, fever, 
and cachexia. The vagina is eventually filled with ulcerating masses and 
there are bulky extensions into the pelvis, and occasionally to regional 
lymph-nodes, rarely to skin or lung (Kalustow, Rosthorn). In Demme's 
case a benign polyp after five years suddenly became malignant. The usual 



histology is that of a large spindle-cell sarcoma, with many blood- and lymph- 
vessels, myxomatous tendencies, and areas of striated muscle. 

Pfannenstiel beheved he could trace the development of muscle-cells 
from the spindle-cells. KoHsko found muscle-tissue in all of three cases and 
believed it to be present in all cases. Islands of cartilage have been observed 
in the primary tumors (Rein, Pernici) or only in recurrences. The remark- 
able clinical characters of this tumor render it a very well-defined disease. 
The early appearance and complex histological structure of the tumors 
show that their origin must be referred to some embryogenic disturbance of 
the cervix and vagina, but the nature of this disturbance has never' been 
clearly defined. 

In the testis rhabdomyoma occurs in the form of nodular tumors as 
large as a walnut or a child's head. They may replace the entire organ or 
displace it to one side, or be found outside the tunica albuginea. In several 
cases the muscle-tissue was associated with epithelial structures, and one 
of Ribbert's cases was combined with carcinoma and sarcoma. Neumann 
and Ribbert found many spindle- and round-cells which were not readily 




Fig. 55. — Structure of congenital rhabdomyoma of heart. (From a section of Wolbach's 


identified as myomatous, and Neumann suggests that some spindle-cell 
sarcomas regarded as fibroblastic may really be of myogenous origin. 
Benenati describes a large rhabdomyoma of an undescended testis with 
round-cell metastases in the regional lymph-nodes. The tumor contained 
spindle-cell and round-cell areas as well as striped muscle. Stoerk observed 
a very malignant case with myomatous metastases in distant lymph-nodes. 
Wood found cartilage and epithelial cysts as well as muscle. All these 
observations lead one to accept the view that rhabdomyoma testis occurs 
only as a one-sided development of a teratoma. Rokitansky and Neumann 
derived their tumors from the gubernaculum of Hunter, but these growths 
were attached to the lower pole of the tunica and were not strictly tumors 
of the testis. 

Congenital rhabdomyoma of the heart is a very characteristic condition 



of which twelve cases have been reported (Wolbach, Lit.)- They occur 
as multiple sharply circumscribed areas or tumors, lying within the wall 
or projecting internally or externally. Ponfick pointed out that they are 
associated with diffuse sclerosis of the cerebral cortex and with disturbances 
of nutrition. Cesaris-Demel found also nodules in the kidney composed of 
embryogenic renal tissue, and Wolbach observed multiple neuroglia rests 
in the spinal meninges. 

The structure of the tumor presents a peculiar spongy tissue in which 
it is difficult to make out the relation of certain large vacuoles to cells and 
fibrils. The vacuoles have been shown to he in the very large cells of which 
the tumor is composed (Seifert, Wolbach). The cells are large rounded 
masses containing large nuclei and nucleoli, surrounded by partly striated 

Fig. 56.; — Rhabdomyosarcoma, of voluntary muscle following fracture of femur. {After 


cytoplasm which extends in many fibrils between the vacuoles to the cell 
border. They resemble early embryonal heart muscle-cells (Kolisko), and 
similar cells may be found in the fetal myocardium (Wolbach). The cross 
striations are formed of fuchsinophile granules grouped in the sarcous 
elements, and the long striations are made up of alternating sarcous elements 
and basophile fibrillary material. Either of these elements may be in excess. 
The resemblance of the cells to Purkinje's cells and those of the conducting 
bundles of the heart is mentioned by Knox and Shorer. That the tumors 
arise from an embryogenic disturbance in the structure of the heart is 
obvious. Kolisko mentions the early and extensive changes in the mor- 
phology of the heart as favoring errors in development. Bonome traces 
the pathogenesis to fetal malnutrition resulting in cerebral atrophy and 
sclerosis, and to fibrous overgrowth in the heart which separates nests of 
embryonal cells from which the tumors develop. 

MYOMA 217 

Miscellaneous Rhabdomyomas. — Tumors containing striated muscle 
have been observed in isolated cases in many other organs and tissues; 
in the esophagus (Wolfensberger, Glinski); in the stomach (Brodowski), 
in the tongue (Pende), parotid gland (Prudden); in the breast (Billroth); 
in the prostate (E. Kaufmann); in place of the left lung (Helbing). 
Rhabdomyomas of the skeletal muscles in various parts of the body, orbit, 
neck, pelvis, buttocks, and different portions of the extremities have been 
collected by Fujinami from reports of various authors, and Benenati gives 
an extensive chronological list of cases. Among the regions affected are 
the spinal vertebrae and pectoralis major (Buhl), upper arm (Billroth), tibia 
(Lambl), thigh (Fujinami), and nose (Erdmann). 

Of 1 6 cases of rhabdomyoma derived from adult voluntary striated muscle 
in which the cells were said to have retained cross striation, reported up 
to 1913, Kuttner rejects six and regards the others as questionable. Tumors 
composed of more anaplastic cells in which cross striation was lost and whose 
origin from muscle-cells was established on other grounds, are very rarely 
reported. Miiller has reported from this laboratory one such case following 
repeated trauma of a long ununited fracture of the femur. The tumor was 
composed of large acidophile spindle-cells, some of which approached the 
dimensions of muscle-cells. A few cases of similar origin are reported. 


' '• Angioma is a tumor composed of newly formed vessels. Both blood- and 
lymph-vessels are subject to neoplastic growth, giving the two classes of tu- 
mors, hemangioma and lymphangioma. 


'■ The determination of the scope of neoplastic processes affecting blood- 
vessels presents unusual difficulties. Vessels occupy an altruistic^ position 

Fig. 57. — Plexiform angioma of skin. Slowly growing and benign. 

in the physiology of organs, subordinating themselves to more specialized 
structures. Possessing less natural autonomy they mayHbe expected to 
display pronounced neoplastic properties. Thus practically normal blood-- 
vessels support highly malignant tumor-tissue. The supply of vessels in a 
tissue or organ varies extremely in both normal and pathological conditions. 
Not being composed of a simple tissue tumors of vessels must be organoid 
in character. The growth of vessels is markedly influenced by the element 
of mechanical pressure of the circulation, which is absent in other tumors. 
Finally the nutrition of vessels, especially of those subject to tumor growth, 
is provided in a different manner from that existing in other tissues. Thus 
the physiology of vessels necessitates special standards in the interpretation 



of tumor processes, and this peculiarity has led to much difference of opinion 
as to what constitutes an angioma. 

From the angiomas may at once be excluded several processes marked by 
overgrowth of vessels. 

Hypertrophic graiHolation tissue presents an extensive newgrowth of 
vessels_ in sinuses, abscess walls, suppurating cysts, and in closed tissues, 
with circumscribed enlargement of tissue. Yet it is self-limiting, regresses 
upon removal of the irritant, and the structure of the vessels shows absence 
of the histological signs of a neoplasm. In rare instances, however, without 
cUnical data it may be difficult to separate this process from angioma, but a 
definite relation between chronic granulation tissue and any of the charac- 
teristic vascular tumors can rarely be established. On the other hand there 
are many indications that certain sarcomas arise from granulation tissue. 

Chronic varicosities are established in veins and arteries in several con- 
ditions which do not constitute tumors, as varicose veins in the limbs, sper- 
matic cord, and broad hgament, and in diffuse aneurisms which follow trauma. 
In hemorrhoids the varicosities are the result of venous stasis and chronic 
inflammation. In inflammatory polyps extensive development of veins 
often results from venous stasis. 

Excessive development of blood-vessels occurs in many tumors as a 
result of venous stasis or unusual demands for nutrition. Very often such 
vessels present a normal or nearly normal structure and do not exhibit neo- 
plastic growth. Such tumors are not true angiomas but their vascularity 
may be indicated by the term " angiomatosum," as fibroma angiomatosum. 
The demands for nutrition in rapidly growing cellular tumors may produce 
a type of growth which is composed chiefly of blood-vessels sheathed by 
masses of tumor-cells, and which is often called perithelioma, or telangiec- 
tatic sarcoma. This rather characteristic structure is not confined to any 
one class of tumors and it is not essentially a tumor of blood-vessels. 

Over against these processes in which there is excessive development of 
more or less normal vessels stands a large class of tumors in which a neo- 
plastic process affects the walls of vessels and usually also the supporting 
connective tissue, and these constitute the true angiomas. 

In general angiomas are of congenital or early development, slow growth 
and benign course, and they occur under rather characteristic clinical 

Hemangioma Simplex. Vascular Nevus (Telangiectasis) . — This type of 
angioma occurs in several forms. 

Nevus vinosus, or .the port-wine stain, consists of a circumscribed or 
diffuse dilatation and newgrowth of superficial capillaries and venules of the 
derma over which the epidermis is usually thin and delicate. The dilated 
capillaries lie immediately beneath the epidermis so that the color of the blood 
is imparted to the skin. They are usually congenital and the face is the chief 
locality affected. A similar process affects the subcutaneous tissue and fat, 
where it fails to yield a discoloration of the skin. According to Unna 
nevus vinosus is to be separated from the true angiomas of the skin, since it 
consists of a simple telangiectasis of venous capillaries and does not progress 
after birth. Nevertheless it and the deeper congenital nevi have often been 
the source of extensive plexiform and cavernous angiomas. It may be classed 
with Albrecht's hamartomas. 

Plexiform angioma consists of a newgrowth of dilated capillaries in which 
the length of the vessels is increased but the number of new cells is not in 
great excess. Small veins and arterioles are also involved. This process 
causes a definite enlargement and tumor of the subcutaneous tissue, with 



flat or warty projections of the skin. Circumscribed tumors of this class 
occur in the skin of the face and especially in the eyelids of young children. 
In a few cases multiple and diffuse tumors with considerable newgrowth 
of connective tissue have occurred, chiefly in the arms and face of children 
and adults, and have been described under the term elephantiasis hemangio- 
ma tosa, or nevus vase, mollusciformis (Seifert, Fox, Kaposi, Unna). In 
Jackson's case the process followed the course of the facial nerve. The plexi- 
form angioma or deep nevus is located in the derma or subcutaneous tissue 
or in the fat tissue and commonly extends from one to the other. It may 
even invade the muscles and bones and eventually in-volve considerable areas 

Fig. s8. — Simple cellular angioma, or angio-endothelioma. 

and even a large portion of a limb. Occurring in embryonal fissures in face, 
cheek, lip and neck, it has been called ^^SMz-aZ angioma. 

Angioma often begins in the vessels about the sweat-glands which may 
become atrophic and fibrosed or hypertrophic. An imperfect formation of 
lobules may be observed in relation to the cutaneous glands, or from natural 
anatomical septa in the fat tissue. The new vessels have the characters of 
arterioles rather than venules, but many sizes and types of vessels may occur 
in the same tumor. Cellular hyperplasia in the vessel-walls is constant. 
The endothehum is much increased and appears in one or more layers. 
It is sometimes exfohated or degenerated, and dilated vessels may be lined by 
giant-cells. Elastin is deficient or absent, and muscular tissue is usually 

Angioma of the muscles occurs in subjects under 30 years of age, chiefly 
in the triceps femoris and forearm, occasionally in other regions. It produces 


moderate enlargement of the part with turgescence and impairment of func- 
tion. Muscatello describes four histological varieties: (i) Capillary and 
progressive; (2) arterial, (,3) venous, (4) cavernous. Pupovac observed 
multiplication of lymph- as well as blood-vessels. In a series of 46 cases 
tabulated by Sutter the above histological varieties were variously combined. 
The process usually began with a proUferation of muscle-cells in the walls 
of venules and in several cases there were polypoid myomatous outgrowths 
into the lumen and fusion of contiguous vessel-walls with subsequent dila- 
tation of channels. The atrophic muscle-fibers were replaced by fat-cells. 
Angioma of muscles may extend to other tissues. I have examined a plexi- 
form angioma involving the subcutaneous tissue, fat, and nearly all the mus- 
cles of the forearm of a young girl. 

Hemangioma hypertropkicum (Ziegler) is a cellular form of capillary 
angioma occurring chiefly in the skin. Nauwerck described a very similar 
tumor occurring in the femur. It consists of a large number of small vessels 
lined by hypertrophic and neoplastic endothehum. The vessels usually 
maintain a scanty lumen but the proliferation of endothelium may obliterate 
the lumen aiid yield compact groups of cells. In this form the tumor is vir- 
tually an endothelioma and in this and the transitional forms it may be des- 
ignated as hemangio-endothelioma. Distention of the cellular vessels may also 
occur giving a cavernous variety of the tumor. In man}? simple angiomas 
especially in the growing edges and in fat lobules, the structure is that of 
hemangioma hypertrophicum. Pure tumors of this type are usually pro- 
gressive and if very cellular may exhibit local malignancy. 

Histogenesis. — Ribbert has analyzed in detail the origin and growth of 
simple angioma. By means of interstitial injection he finds that the vessels 
have few or no lateral anastomoses, while the injection mass passes freely 
into the efferent artery and afferent veins but not into the tissue surrounding 
the tumor. This result indicates that the tumor process resides in an iso- 
lated segment of the vessel-walls and produces elongation and varicosities 
in a more or less closed territory without gradual involvement of surrounding 
vessels. Where the tumor forms a new lobule or invades fat or other tis- 
sues it is not by extension of the tumor process to the healthy vessels of the 
new area but by the projection of new vessels which grow out from the tumor 
while the vessels of the invaded part are compressed and occluded. The mi- 
croscopical structure shows that the process affects both the walls of the ves- 
sel and the supporting connective tissue. In the invaded fat lobules Rib- 
bert finds isolated tumor-vessels connected by long strands with the main 
tumor, while between the fat-cells parallel rows of endothelial cells appear 
which later become thickened and canalized. These new structures seem to 
have no connection with the normal vessels of the invaded tissue. 

Thoma refers much of the growth of angioma to mechanical factors. 
Increase of blood-pressure and loss of support to vessel-walls from changes 
in the surrounding tissue tend to excite new growth of vessels, while increased 
rapidity of flow favors elongation and dilatation of the wall. Many authors 
(Rokitansky, Borst) consider that many simple so-called angiomas represent 
simple hypertrophy of vascular segments without neoplastic overgrowth. 
These factors are doubtless important in determining the course of angiomas 
but they cannot account for their origin, which must be referred to a develop- 
mental anomaly in the structure of certain vascular segments which do not 
fit into the circulatory system, and which retain embryonal characters. 
The congenital origin of the great majority of angiomas speaks strongly in 
favor of a tissue predisposition as a prominent factor in their genesis. 

Virchow believed that angiomas result from the action of local irritation 



on imperfectly formed vessels, as those in embryonal fissures. Unna was 
led to believe that nevus vinosus especially occurs in areas which have been 
subjected to abnormal pressure during fetal life. The relation to nerve- 
trunks has suggested to many a neurotic theory of origin. In not a few 
cases a traumatic origin is clearly indicated (Lowenthal, Lit.). 

Cavernous Angioma. — When the vascular channels are widely dilated 
and the connective tissue septa are thin, the angioma is designated as cavern- 
ous. This process is of frequent occurrence in many situations and in nearly 
all tissues and organs. In the skin cavernous angiomas produce circum- 
scribed or diffuse, flat or elevated lesions, involving the derma and subcuta- 
neous tissue. If superficial they are dark red in color. If covered by cor- 

FiG. 59. — Metastasizing angioma, secondary nodules in lung. {Borrmann.) 

rugated and thickened epidermis, they become warty. Changes in the 
circulation ailect both the size and color of the tumors. Many of them are 
distinctly erectile and pulsating. Thrombosis and the formation of cal- 
cific phleboHths occur in dilated sinuses. The smaller tumors are usually 
encapsulated and stationary but there are many cases of diffuse cavernous 
angioma in which a capsule is missing and the limits of the process gradu- 
ally extend over many years. 

The cavernous angioma first appears as a circumscribed tumor developing 
often on the basis of a congenital nevus and tends to enlarge steadily over a 
period of many years. They may be observed at birth or appear at any age. 
When allowed to progress they may attain very large dimensions, and 
successively invade neighboring tissues and organs. In some of the older 



cases, extraordinary results were produced. Gascoyen observed an angioma 
of the parotid which progressed for many years, produced a large polypoid 
tumor externally while a pharyngeal portion eventually caused death by 
suffocation. Autopsy showed several nevi of the intestinal serosa and 
mucosa and one in the liver. Cruveilhier described a cavernous angioma 
involving the skin, muscles, tendons, synovial membranes, nerves, and 
periosteum of nearly the whole arm of a 65-year-old hemiplegic. Falkowski 
reports peculiar cavernomas of liver and spleen' and angiomas of skin in an 

The subcutaneous angiomas may gradually involve the skin and the 
deep fascias, and some which first appear subcutaneously originate in much 
deeper tissues, or they establish wide communications with deep venous 

Multiplicity of cavernous angioma is a prominent feature in some cases. 
The minute senile angiomas of the skin usually appear in considerable num- 

FiG. 60. — structure of metastasizing angioma. (Borrmann.) 

bers. Large multiple cavernomas of widely separated regions occurred in a 
case of Hildebrand's. The smaller cavernous angiomas are often multiple, 
as many as forty or even a hundred having been observed in one subject 
(Esmarch, Schuh). 

Sharply contrasted with the multiple benign cavernous angiomas is a 
group of metastasizing cavernous or more cellular angiomas which exhibit 
certain peculiar features of malignancy, and are eventually fatal, chiefly 
through internal hemorrhage and anemia. Here belongs the remarkable 
case observed by Borrmann in which angioma of the skin of the breast in a 
subject of 23 years recurred repeatedly after operation and finally proved 
fatal with numerous secondary tumors in both lungs. The original tumor 
had the structure of simple angioma but the secondary growths were 
more cellular. 

In a case observed by the writer the breast was greatly enlarged by a 
bulky cavernous angioma, several tumors appeared on the skin and mucous 



membranes, and there were evidences of pulmonary involvement. The 
structure was very similar to that in Borrmann's case. 

In Shennan's remarkable case, of six years' duration, marked by numerous 
hemoptyses, there were found at autopsy cavernous angiomas involving 
the whole of the spleen, much of the lungs, the thymus and mediastinal nodes, 
liver and bone-marrow,, while many miliary angiomas occurred in other 
tissues. The structure was generally not malignant in appearance, but 
the ^ small tumors and some of the larger were doubtless metastatic, and 
exhibited a local invasive tendency. 

Fig. 6i. 

-Metastasizing angioma of left breast. Note tumors of jaw, neck, under ear, and 
at right axiUa. There were tumors also in pharynx and lungs. 

The course of cavernous angioma is usually slowly progressive. Begin- 
ning in congenital nevi or in deep tissues, at any period of life, they slowly 
enlarge by distention of the original vessels and by formation of new vessels. 
The growth is sometimes accelerated at the menstrual period and during 
gestation. Free anastomoses with large arteries or veins may become estab- 
Ushed and severe or fatal hemorrhage may occur. A capsule forms about 
many tumors which are then apt to remain stationary. Spontaneous 
regression may result from contraction of the capsule, thrombosis, inflam- 
mation, or ulceration, or a cure may be affected by continuous pressure, 
ligature of vessels, or excision. Very slight interference is sometimes 
sufficient. Intercurrent diseases or cachexia may initiate the regression. 



Recurrence after operation has often followed incomplete removal. A 
certain local malignancy is also exhibited by tumors which extend from one 
tissue to another, leading even to the erosion of bones, and occasionally 
pressure symptoms become serious. 

The structure of cavernous angioma presents chiefly a series of anastomos- 
ing vascular, channels inclosed by thin septa. The appearance on section 
may be roughly compared to that of a sponge. Yet in less advanced cases 
it is possible to recognize simple spherical sacculations, and tubular dilata- 
tions, all connected with one main afferent and one efferent vessel. Large 
cystic dilatations may occur, filled with blood or serous fluid. The walls 
exhibit the structure of venules or less often of arterioles. Hyperplasia 
of cells may lead to nodular growths of new tissue projecting into the lumen. 

Fig. 62. — Cavernous angioma of liver. 

Many authors have described the budding of new vessels on the advancing 
edges of cavernous angioma. In progressing cases certain areas of the tumor 
may show the numerous cellular vessels of simple or plexiform angioma, 
so that it is not always possible to separate cavernous from simple angioma. 

V. Reckhnghausen and others locate the earliest stages of^ the cavernous 
angioma in the walls of the veins, while many believe the capillaries are first 
affected (Borst) . I have drawn the impression that either veins or capillaries 
may be involved in different cases. 

Histogenesis. — In the development of cavernoma Rindfleisch and Borst 
attribute chief importance to a fibrocellular growth in and about the walls 
of the capillaries. The retraction of this new tissue tends to shorten the 
vessel, dilatation resulting from mechanical pressure. The same relation 
between connective-tissue growth and vascular dilatation is observed in 
cavernous fibromas, so that the cavernoma has been hkened to a fibroma 



with excessive development of vessels. Loss of muscular and elastic tissue 
in the new or altered vessels must also greatly favor the dilatation. In 
regressing cases a fibrous capsule forms or fibrous areas appear in the tumor, 
the sinus walls thicken and contract, and eventually only scar tissue remains. 

The question of the neoplastic nature of cavernous angioma has been 
extensively discussed. When one examines a stationary cavernoma of the 
liver it is difficult to recognize any definite feature of a tumor and many 
have assumed that these and a considerable proportion of other locahzed 
dilatations of vessels should be separated from the angiomas and referred to 
mechanical factors. There is no doubt that simple varices may simulate 
the structure of cavernous angiomas and have often been classed with this 
tumor. The gross dissection of certain angiomas strongly suggests that 
simple saccular and tubular dilatations of preexisting veins will explain 
their origin. Yet Virchow traced the earliest stages of cavernoma of liver 
to islands of proliferating connective tissue surrounding cellular capillaries, 
and Ribbert finds on the edges of cavernoma new vessels which communicate 
freely with those of the tumor but imperfectly with those of the surrounding 
tissue. Local varices can usually be distinctly separated from angioma in 
the same tissue. The usual origin of cutaneous angioma is from congenital 
simple or plexiform angioma which is admittedly a neoplasm. It is thus 
apparent that in the growth of cavernous angioma there are essential factors 
other than mechanical dilatation of vessels, and it is most reasonable to 
regard these factors as partaking of the neoplastic order. Albrecht regards 
cavernoma of liver as an ill-fitted but practically normal segment of tissue 
which possess a limited capacity for aberrant growth. For this and many 
other benign tumors he introduced the term hamartoma. 

The clinical features of cavernous angioma fall into several rather well- 
defined groups. The cutaneous tumors form the most numerous group. 
They are located in the derma or subcutaneous tissues, chiefly in regions 
where the skin is loose. While no region of the body escapes, the chief 
locations are the face, scalp, labia, scrotum, prepuce, extremities, and folds 
of the knees, axilla, and buttocks. The influence of embryonal fissures 
noted in the occurrence of simple angioma is observed also with the cavernous 
type. The wide variations in the origin, growth, size, number, and complica- 
tions of these tumors is elaborately set forth in the detailed reports of cases 
collected by Virchow. 

Submucous cavernomas occur chiefly in the buccal region. The gum 
is the seat of a common cavernous tumor which may follow violent extraction 
of molar teeth. It usually remains of moderate dimensions but may gradu- 
ally extend until it involves a considerable portion of the alveolar tissues 
and establishes connections with large veins. At this stage its extirpation 
is difiicult and has produced fatal hemorrhage. 

In the tongue cavernous angioma arises usually in the tip of the organ 
and may extend until it produces an erectile tumor of large dimensions 
(Reiche) . Both lymph- and blood-vessels participate in the process in certain 
cases of angiomatous macroglossia (Wegner). A form of angiomatous 
ranula has been described by Delbeau. The lips are frequent seats of con- 
genital nevi and of cavernous angiomas. 

Orbital angiomas affect either eyelid and may involve the conjunctiva. 
In the retroorbital fat a few diffuse and circumscribed cavernous angiomas 
have been described (Morton, Schuh, di Ricci). These tumors may extend 
along the optic nerve into the eyeball (Quackenboss). 

Angioma of bone is rare, although overgrowth of vessels is a prominent 
feature of many malignant bone tumors, and bone-tissues are often eroded 


by extensions from angioma of adjacent tissues. The extensively cavernous 
tumors pf the long bones, especially of humerus and tibia, must be classed 
histologically as sarcomas, although some of them may be successfully treated 
by curettage or ligation of the main vessels. True cavernous angioma may 
arise in the periosteum or in the marrow (Virchow, Kauffmann). They in- 
vade the outer layers or cause absorption of the shaft. Extensive angiomas 
of the skull have been observed by Kauffmann and by Schoene, and of the 
vertebras by Gerhardt, Kauffmann and Muthmann. I have studied one 
cavernous angioma of the head of the humerus which appeared beneath a 
thin shell of bone two years after trauma. In a case of slowly progressive 
paraplegia I found a cavernous angioma of a dorsal vertebra, which permitted 
collapse of the body and gave rise to a protruding tumor which compressed 
the cord. 

In the glands angioma occurred in the parotid in the case of Gascoyen, 
and in several cases the breast has been extensively transformed into caver- 
nous tissue by tumors arising in the fat (Virchow, Image, Borst, Hake). 

In the internal organs cavernous angioma occurs chiefly in the liver, 
rarely in the spleen (Albrecht), kidney, and uterus (Virchow). Dowd has 
collected 13 cases of angioma of the spleen, illustrating small tumors found at 
autopsy, very large growths removed at operation, malignant tumors produc- 
ing metastases, multiple benign tumors in spleen, liver, omentum, and skin, 
and a progressive anemia accompanying the large growths. 

The common cavernoma of the liver is of much theoretical interest but 
rarely produces symptoms. They are single or multiple, as small as a pea 
or as large as a child's head (Ribbert), and usually lie just beneath or project- 
ing from the surface. They are often associated with angioma of other or- 
gans. Their frequency increases with the age of the subjects but congenital 
cases rarely occur. Veeder and Austin describe a remarkably extensive 
multiple congenital hemangioma of the liver. Virchow believed that they 
might spontaneously disappear, leaving scars. The structure shows central 
sinuses separated by thin fibrous or cellular septa which may contain 
islands of liver cells. On the edges are smaller vessels communicating with 
the sinuses. The tumor is sharply marked off from the parenchyma. Rib- 
bert finds that interstitial injections of the tumor or of the surrounding 
parenchyma do not pass from one tissue to the other, the tumor-vessels 
showing a marked independence of the normal vessels of the organ. Yet 
the tumor may be filled by injections through the portal or hepatic veins or 
hepatic artery (Virchow). 

The origin and nature of the cavernoma of the liver has long been a 
subject of discussion, the later phases of which have been maintained by 
Ribbert and by Schmieden. On the whole the argument seems to favor the 
contention of Ribbert, that these growths are partial neoplasms originating 
from embryogenic disturbances, through which a displaced segment of the 
organ comes to possess a limited power of aberrant growth. As already 
stated, Virchow traced their inception in areas of proliferating vascular con- 
nective tissue. The progressive dilatation of vessels naturally results from ab- 
sence of muscular and elastic tissue in the walls. 

In the ovary Gottschalk describes a diffuse bilateral cavernous transfor- 
mation. Orth observed a remarkable case of multiple congenital hemangi- 
omas in both ovaries, skin, and other organs of a child. 

Angioma of the brain occurs in two main forms, as (i) a more or less 
diffuse varicosity of the vessels of the meninges, and (2) a true angioma, sim- 
ple or cavernous, of the brain tissue. La Villette has collected 18 cases of 
both types in \-arious situations. Several cases have occurred in the region 


of the corpus striatum (Virchow, Luschka). They are probably of con- 
genital origin, and in Heboid's case the intracranial was associated with cu- 
taneous angiomas. They are of slow growth and usually produce only gen- 
eral symptoms. Wergman's cavernous angioma of the left cerebellar 
pedicle led to fatal hemorrhage. Orbison reported a racemose angioma of 
cerebral pia invading the brain tissue and giving a long history of epi- 
lepsy. Several cases of angioma in the substance of the pons are recorded by 
LaFora and by Enders. One of LaFora's cases terminated after severe gen- 
eral and focal symptoms lasting one month. The structure is that of sim- 
ple or cavernous angioma. 

Angioma arteriale racemosum consists of a dilatation and complex inter- 
twining of many new formed and altered vessels of small caliber with sub- 
sequent involvement of normal vessels (Deetz). The condition occurs in 
infants or adults, and of 87 cases collected by Schuck 84 occurred on the 
head, a few on the extremities. They usually arise in close connection with 
the large carotid artery. The external appearance was likened by Virchow to 
a pulsating mass of earthworms. Arising externally they extend over the 
neck and scalp, erode the skull and penetrate the cranium, and in several 
cases they have involved the cerebral and meningeal vessels. The walls of 
the vessels show a variety of changes, as fatty degeneration and loss of 
muscle- tissue of the media (Heine), hypertrophy of media (Lablee) hyper- 
trophy of interna (Kretchmann) , or hypertrophy of all coats.' Emanuel 
described leiomyomatous outgrowths of the media with degeneration, calcifi- 
cation and formation of aneurisms. Deetz found uniform hypertrophy of 
small arterioles and newgrowth of cellular capillaries. The exact position 
of this process is difficult to determine but most authors assume that it 
arises from a congenital tissue abnormality and that it belongs in the class 
of partial tumors. 


Lymphangioma is a tumor composed of lymph-vessels. 

It is therefore an organoid structure consisting of endothelial cells and 
supporting connective tissue, both of which are involved in the neoplastic 
process. Lymph-nodules or foci of round-cells are often present to complete 
the parallel with a lymphatic structure. Lymphangioma has also been 
associated with minute islands of lipoma, proliferating smooth muscle-cells 
may be found in the septa, and Ritschl found an island of cartilage in an 
intermuscular lymphangioma, observations which suggest an embryogenic 
origin. New formed lymph-vessels are also present in many benign and ma- 
lignant tumors, especially with endothelioma and sarcoma. 

The clinical conditions which fall in the general class of lymphangioma in- 
clude a variety of slowly growing, usually congenital, single or multiple tu- 
mors of the skin, subcutaneous tissues, deep areolar tissues and muscles, of 
the neck, trunk, lips, tongue7 eye and orbit, and mediastinal and retroperi- 
toneal regions. In the pathogenesis of these conditions many factors are 
concerned and it is thus more difficult to distinguish between lymphangioma 
and lymphangiectasis than between true and spurious hemangioma. 

Ribbert's definition of lymphangioma as an isolated group of vessels 
growing from their own resources and more or less disconnected from sur- 
rounding channels can seldom be directly applied. In some typical examples 
of lymphangioma the tumor has been found freely accessible to injections 
through afferent and efferent vessels (Langhans). 

Proliferating buds of endothelium forming new lymph-vessels have been 
fully recognized in certain cases by Nasse, Schmidt, and Borst, but such 


structures are usually missing. Size is not a reliable criterion since in some of 
the large cystic growths the signs of a neoplasm are least distinct. More- 
over, simple occlusion of large lymph paths may, although rarely, be fol- 
lowed by extensive varicosities resembhng lymphangioma. The appearance 
of the stroma sometimes suggests a neoplastic growth but more often it is 
fibrous or comparatively acellular. The general clinical features seem to 
offer an important means of recognition of true lymphangioma, and many 
would include in the class only such processes as show a congenital origin and 
a progressive course, and absence of traumatic and inflammatory factors. 
The histological structure of lymphangioma was first fully described by 
Wegner who recognized three groups of cases, i. Lymphangioma simplex 
consists of an anastomosing network of spaces and vessels of small and 
medium cahber. The septa are either reduced to thin strands of acellular 
connective tissue, or they are thicker and participate in the proliferation. 
The endothelium is flat or cubical and rarely it may appear in multiple 
layers or in the form of projecting buds (Freud weiler). 

2. Lymphangioma cavernosum consists of a system of closed communi- 
cating lymph-spaces supported by thin walls or thicker septa, lined by flat 
endothelium and filled with fluid or coagulated lymph occasionally mixed 
with blood. 

3. Lymphangioma cysioides consists of a congeries of large and small 
cysts lined by flat endothelium and filled with lymph. Some of the cysts 
may be closed and the main lymphatic vessels are usually occluded. 

The great majority of lymphangiomas fall readily among these classes, 
most of them being of the cavernous type. In not a few cases there is new- 
growth of blood-vessels not communicating with the lymphatic system, and 
forming a mixed hemolymphangioma (Nasse, Sutter, Novack). 

In the development of lymphangioma several factors seem to be con- 
cerned. Wegner believed that three modes of origin could be traced, (a) 
passive dilatation with inflammatory hyperplasia of preexisting vessels, 
lymphangiectasis, (6) neoplastic growth of vessels, and (c) heteroplastic for- 
mation of lymph-vessels in granulation tissue. 

{a) Lymph stasis doubtless influences the course of many lymphangiomas, 
especially the large cystic type with occluded vessels, to a less extent the 
saccular ectasias of the cavernous forms, but it has not been accepted as a 
sufficient explanation of any of the characteristic varieties of lymphangioma. 
Ribbert argues that any marked pressure would inhibit the growth of endo- 
thelium, and it is clear that with free afferent and efferent vessels the stasis 
must be slight. Ribbert assumes that lymphatic and other cystic tumors 
develop as a result of a process of growth located in the walls which steadily 
enlarges the circumference of the cysts and prevents the passive accumu- 
lation of fluids. Secondary dilatation of new-formed vessels must be re- 
garded as an important factor in cavernous and cystic lymphangiomas. 
Intermittent attacks of inflammation influence the course and structure of 
many lymphangiomas (Kuttner). 

(6) The evidence of new formation of vessels is conclusive, for in many 
cases have been observed hypertrophic and proliferating endothelium, 
multiple layers of endothelium, and sprouts of endothelium growing out into 
the connective tissue with subsequent canahzation. Borst depicts the for- 
mation of new lymph- vessels in a small fibromatous tumor in fat tissue where 
lymph-vessels are normally scanty. In a type of cutaneous lymphangioma 
the predominance of endothelium has led many to designate the process as 
endothelioma and other types of endothelioma arise from lymphatic en- 
dothehum. The occasional occurrence of definite fibroma and lipoma in 



lymphangiomas illustrates the organoid character of these tumors and 
demonstrates the neoplastic growth of lymph- vessels. 

(c). The heteroplastic formation of lymph- vessels in granulation tissue 
has not been satisfactorily demonstrated. 

In the origin of lymphangioma it must be assumed that there exists a 
local predisposition resulting from an embryogenic disturbance similar to 
that assumed for hemangioma. Of the nature of this disturbance nothing is 
definitely known, but the congenital origin of most lymphangiomas is a strik- 
ing feature in their etiology. A partial isolation of a segment of lymph- 
vessels with imperfect development and retention of abnormal powers of 
growth may be supposed to exist. The occurrence of cartilage in one re- 
ported lymphangioma indicates that the isolation may be complete. I have 
seen a circumscribed lymphangioma 2 cm. in diameter attached to the sper- 

FiG. 63. — Structure of a congenital angiolymphoma occurring in skin of cheek of an infant. 

matic cord. The vessels were lined by large cuboidal endothelium and sup- 
ported by embryonal connective tissue. In a congenital flat tumor of the 
subcutaneous tissue of the cheek I found many lymph-follicles and diffuse 
lymphoid tissue in the septa of a cavernous lymphangioma. The structure 
suggested an abortive attempt at the formation of a lymph-node. 

The clinical forms of lymphangioma are numerous and difficult to classify. 

Simple lymphangioma, consisting of moderately dilated and slightly 
hyperplastic vessels in which a neoplastic element is either wanting or feebly 
developed, occurs in many situations, chiefly in the skin. In elephantiasis 
dilated lymph-vessels contribute largely to the increased bulk of the tissues. 
In the lymph-scrotum of filariasis the swelling is referable to lymphangiec- 
tasis and edema, and may be traced to occlusion of lymphatic trunks by 
the parent worms. The process commonly involves the bladder with chy- 



luria and the presence of filaria embryos in the urine. The tissues show only 
inflammatory reaction and the process is not a tumor. 

In the skin of face and neck congenital lymphangiectasis produces small 
or large flat or wart-like prominences which after trauma may exude serous 
fluid. They consist of dilated lymphatics of the derma. The endothelium 
is slightly hyperplastic, but their chief neoplastic character is the occasional 
relation to embryonal fissures and their congenital origin. In a case of Haug's 
the tongue was involved and there was a congenital auricular fistula. 

In lymphangioma cutis circumscripta, the skin of face, chest or extremities 
is the seat of numerous small projecting translucent vesicopapules which on 


Fig. 63a. — Superficial diffuse lymphangioma of entire tongue, existing since infancy in 

a boy of 19 years. 

section prove to consist of many small cellular lymph-vessels. The endothe- 
lium is hypertrophic and proliferating, and the channels are smaU and filled 
with fluid or hyaline masses. There is extensive new formation of imperfect 
lymph-vessels and the process is a true neoplasm (Fox, Hutchinson, Noyes, 
Torok, Freudwefler, Waelsch). 

Lymphangioma tuberosum multiplex was described by Kaposi as a small 
superficial multiple tumor of ceUular lymph-vessels. Beneke's case falls 
in that histological group. The nature of other cases in dermatological 
hterature is not clear (Crocker, Perry, Kromayer). 

In the eye lymphangiectasia; of congenital or inflammatory origin occur 
in the conjunctiva. They may reach considerable dimensions and occasion- 
aOy there "are evidences of considerable new formation of vessels (Steudener, 
Bull, Wintersteiner, Sachs). 


Lymphangioma cavernosum occurs chiefly in the skin but also in the 
intermuscular septa and in the mucous membranes. The dilated vessels 
are filled with coagulable lymph often mixed with blood, and the contents may 
be thrombosed, or hyaline, or extensively calcific. On analysis the aspirated 
fluid shows a considerable content of blood-proteins and salts, many lympho- 
cytes and exfoliated endothelial cells and usually cholesterin crystals. It lacks 
the digestive ferments of pancreatic cysts (Sick, Bryck, Zeyneck) . The septa 
are thin and acellular or thicker and composed of proliferating connective 
tissue, containing round-cells, lymph-nodules, and often much smooth muscle- 
tissue. Newgrowth of small vessels is often associated with the larger ones. 
The tumors are small and circumscribed or extend diffusely over considerable 

In the skin elephantiasis lymphangiectatica may exhibit extensive dilata- 
tion of old and new formed lymph-vessels. 

The most frequent forms of cavernous lymphangioma occur in rhe lip, 
cheek, and tongue and are designated as macrocheilia, macromelia, and 
macroglossia. They are usually congenital and produce irregular and some- 
times extensive enlargement of the tissues. The pharynx and larynx may 
also be involved (Suchstorff, Nasse). 

Chylangioma of the mesentery is a cavernous lymphangioma containing 
milky fluid which arises from congenital or acquired obstruction to the lacteal 
vessels (Kruse, Schmidt, Ritter) . Multiple dilatations of the submucous 
lymphatics of the intestine were observed by Krauss, and extensive grades 
of this condition have been reported by Blatteis, Thalheimer, and Takano. 

Cavernous lymphangioma occasionally occurs in the conjunctiva, eye- 
lids, and orbit (Michel, Steudener, Sachs). Klien described a case occurring 
in the vagina. 

Extensive cavernous lymphangioma of the foot resembling elephantiasis, 
multiple lymphangioma of the finger, hands and pleura, and hyperplasia of 
many lymph-nodes, were associated in a remarkable case reported by Bryck. 

Lymphangioma cysticum is a multilocular cystic tumor which occurs 
chiefly in the neck and sacral region. They are usually of congenital origin 
and exist at birth or shortly develop into tumors of considerable bulk and 
wide extent. 

Hygroma cysticum colli is a lymphangioma which usually arises in the 
submaxillary region and ramifies upward toward the parotid and ear, inward 
toward the median line,. and downward to the supraclavicular fossa and even 
into the pleura and mediastinum (Nasse, Suckstorff). One class of these 
tumors occurs deep in the neck, reaches below the sternal notch and seems 
to be connected with the thyroid (Otto). In one case which came to my 
notice the cyst lay close to the trachea and reached from the sternal notch to 
the inferior thyroid artery. Lymphangioma of the neck must be distinguished 
from hydrocele colli, which is a simple dilatation in a branchial cleft, 
lined by epithelium and walled by fibrous tissue containing lymphoid tissue 
(Frobenius). The axilla and thoracic wall are occasionally the seat of cystic 
lymphangioma (Nasse). Bilateral congenital cystic lymphangioma of the 
back of the neck has been described by Frobenius. 

Retroperitoneal and mesenteric lymphangioma occurs in children and adults 
as a multilocular cavernous and cystic tumor, originating along the spinal 
column and ramifying into the pelvis behind the kidney or colon, upward to 
the liver, spleen and pancreas, and into the mesentery or omentum (Takano, 
Lit.). The receptaculum chyli and thoracic duct have been found unaffected 
by the process (Sick, Smoler, Schwarzenberger) . Lion has reported a case 
of lymph-cyst of the broad ligament. In a case of the writer's the broad 



ligaments were chiefly involved with extensions throughout the pelvis and 
along the lumbar spine. 

The very early stages and exact origin of these tumors have not been 
demonstrated, and as Hedinger points out, they require differential diagnosis 
from a variety of other abdominal cysts. Their neoplastic nature seems 
assured since the walls contain cellular connective tissue, often much smooth 
muscle-tissue, and lymph-follicles. After partial removal the remaining 
portion may rapidly increase in bulk. Along the edges of the growths. 
Sick found proHferating areas of cavernous lymphangioma. He assumes that 
they arise from misplaced and embryonal islands of connective tissue and 
lymph-vessels. In the omentum and mesentery of newborn cats and pigs 
Ranvier has demonstrated such misplaced islands of tissue. 

Fig. 64. — Papillary growth in a cervical lymph-node which was the seat of chronic lymph- 
stasis. Papillary fibro-endothelioma. 

In the wall of the stomach and gastrohepatic ligament, Reimers and 
Tilger have found small lymph-cysts which they referred to obstruction 
of vessels following gastric ulcer. 

Sacral hygroma is a cystic tumor developing in the region of the sacrum 
and usually connected with defects in the spinal canal. Some of them are 
probably aberrant meningoceles. Borst found a large multilocalar cystic 
tumor in front of a defective sacrum and connected by a strand of tissue with 
the spinal canal. Virchow described an external pedunculated hygroma 
attached to the sacrum of a negro infant and containing a lymph-cyst and 
portions of nervous tissue. Attached by a pedicle to the hver of a child 


of five years, Maresch found a large cavernous and cystic lymphangioma. 
In the adrenal and in the uterine musculature, lymph-cysts have been de- 
scribed by Sick. 

In the ovary Kroemer finds lymphangioma of small dimensions not in- 
frequent. Besides telangiectatic fibroma he describes two large cystic 
lymphangiomas, in one of which the endothelial cells werej prominent, 
suggesting the designation fibro-endothelioma. 


General Characters. — Sarcoma is a malignant tumor composed of cells 
of the connective-tissue type. 

This definition is based on the morphology of the tumor-cells and on their 
histogenesis. It accomplishes the main object of separating a large group 
of malignant tumors from the carcinomas, but it overrides certain embryo- 
logical considerations. 

Thus it permits the inclusion of gliosarcomas which are of ectodermal 
origin, and it accepts as sarcoma certain tumors, chiefly of endothehal origin, 
while rejecting others. This position seems justified by the history of the 
development of the conceptions of sarcoma, as ably sketched by Malherbe, 
by certain anatomical and clinical features of the group, and by the peculiar 
part played by the blood-vessels in the growth of many sarcomas. Never- 
theless, it cannot be claimed that the effort to stretch the scope of the term 
sarcoma over an extensive field of neoplasms has been entirely successful. 

The diseases included in this group, such as angiosarcoma, lymphosarcoma, 
and gliosarcoma, are of such varied origin and character that some writers 
have urged the elimination of the term sarcoma. This radical reform has 
not received approval since there are many common features among 
sarcomas and the group as a whole is well-defined from carcinomas. Yet 
future investigation will doubtless reveal many new and more precise facts 
regarding the etiology, conditions of incidence, histogenesis, and clinical 
course which will warrant the recognition of many sarcomas as specific 
pathological entities. 

The original conception of sarcoma as a tumor of fungating, soft or fleshy 
character reflected truly the gross character of most sarcomas, but the 
definition based on histogenesis is much more significant. Even this con- 
ception of sarcoma we believe is destined to be replaced by etiological criteria. 

Since there is often much difficulty in determining the origin of cellular 
tumors the accepted scope of sarcomas has been subject to much revision. 
In recent years, especially, many tumors once interpreted as sarcomas, have 
been proven to be of epithelial origin. Further inroad has resulted from the 
identification as endotheliomas, of many tumors once regarded as sarcoma, 
but it appears unlikely that the separation between the latter groups will 
be rendered complete. 

In fact the finer analysis of the origin and composition of many sarcomas 
reveals a prominent participation of endothelium in many tumors of distinct 
mesoblastic characters. In such cases the characters of the tumor-cells 
rather than their embryonal antecedents should determine the classification. 

Even the spindle-cell sarcomas have not escaped suspicion, especially 
in epithelial organs, since it is a fact too often ignored that squamous and 
glandular epithelium may assume a fusiform shape during the vicissitudes of 
tumor growth. 

Gross Anatomy. — The majority of sarcomas are bulky, soft, and vascular. 
In general, the bulky character results from the principle that sarcomas tend 
to grow expansively upon a scaffolding of new blood-vessels rather than_^to 



infiltrate preexisting tissues as do carcinomas. The soft consistence reflects 
the cellular structure of a rapidly growing tumor. The vascularity reveals 
to the naked eye the prominence of blood-vessels as an essential part of the 

Some specific varieties of sarcoma fail to display the usual features. 
Fibrosarcoma is firm, resistant, and non- vascular ; osteogenic sarcoma may 
be as hard as ivory. In fact, there are so many variations in the appearance 
of sarcoma that general statements are of little value and the gross characters 
are best detailed in connection with specific tumors. As a rule the gross 
appearance depends largely upon the tissue of origin, according to which 
one recognizes fibrosarcoma, chondrosarcoma, osteosarcoma, liposarcoma, 
myxosarcoma, angiosarcoma, myosarcoma, and lymphosarcoma. All adult 
mesoblastic derivatives are represented in the tissues of origin. Gliosarcoma 
occupies an especially peculiar position. While its cells of origin are derived 
from neural epithelium, the form and function of the adult glia-cells, and 
especially the characters of the tumors, do not encourage efforts to reform 
current nomenclature in this group. 

Sarcomas are usually single, but multiple tumors arise in bone-marrow, 
nerve-trunks, and skin. Certain angiosarcomas are multiple or diffuse, and 
some forms represent systemic diseases, as lymphosarcoma and myeloma. 

Characters of Growth. — The growth of sarcomas is usually rapid and 
locally destructive. The abundant blood-supply favors extremely active 
cell-growth so that the tumors soon reach a large size, and their bulk is often 
sharply increased by edema, hemorrhage, or mucoid degeneration. After 
reaching a certain limit the excessive circulation faUs and large areas of 
necrosis form from infarction. The further growth of the tumor then 
depends on its capacity to infiltrate or produce metastases. 

Hence some sarcomas regress after extensive central necrosis and not a 
few are remarkably susceptible to toxic agents or to absorbents such as w-ray 
or iodide of potash. Hemorrhage from the fragile vessels is prominent 
in many vascular tumors. Into the extra vasated blood-mass sarcoma-cells 
are often observed to grow, replacing the clot. 

Following necrosis, sarcomatous tissue is very susceptible to infection 
and many cases terminate with local suppuration and gangrene or with 
general sepsis. Accordingly, fever is often observed throughout the course 
of malignant sarcomas, but in some instances it appears to result from specific 
intoxication and not from infection. 

Most sarcomas grow expansively, the central portions increasing with 
the peripheral. Hence they are rather sharply marked off from the surround- 
ing tissue although rarely presenting a capsule. 

Infiltrative growth on the periphery fuses the tumor with the surround- 
ing tissue so that it is difficult to extirpate and commonly recurs after opera- 
tion. The more malignant growths infiltrate widely through fascial planes, 
along vessels and nerves and directly through the soft tissues. Penetration 
of the skin is followed by more rapid fungating growth with ulceration and 
hemorrhage. Cellular bone-sarcomas rapidly destroy bony tissue and perfor- 
ate joint cavities. To the attack of other sarcomas bone is resistant. I have 
observed very wide infiltration of a spindle-cell periosteal sarcoma which 
produced a large tumor of the ilium and spread beneath the periosteum over 
the entire ilium, pubes, and upper third of the femur. The extensions of 
lymphosarcoma and of myeloma are partly by infiltration and largely by 
systemic development of new tumors. 

Metastases are observed in the advanced stages of most sarcomas and 
it is characteristic of the disease that with rare exceptions the embolic cells 


travel through the blood-vessels. Except in lymphosarcoma, the occm-rence 
of metastases in lymph-nodes justifies the suspicion that the tumor is not a 
sarcoma. The strong tendency of sarcoma-cells to grow upon a framework 
of blood-vessels determines the early invasion of these vessels which occurs 
even in sclerosing osteogenic sarcoma, while the expansive growth, as opposed 
to the infiltration of preexisting tissues in carcinoma, explains the usual 
failure to invade lymphatics. Sarcoma tissue itself is usually devoid of 

Cell-emboH thus pass readily into the vessels and lodge first in the lungs 
which are the chief seat of sarcomatous metastases, and later they reach 
liver, kidney, spleen and other organs and tissues. 

General miliary sarcomatosis may result from discharge of very numerous 
loose cells, but it is much less frequent than general carcinomatosis. Mechan- 
ical pressure of actively growing sarcoma sometimes forces tumor-tissue 
into the blood-vessels and thus leads to metastases when the infiltrating 
power and capacity for independent growth of the cells appears to be limited. 
Thus myxosarcoma of the thigh may be forced into the saphenous and iliac 
veins and form a continuous mass filhng vena cava, pulmonary veins, and 
even reaching the heart. The same extensions of sarcoma of the kidney 
have been observed. 

Some sarcoma-cells seem comparatively resistant to any destructive 
influence that may be exerted by the blood. They have frequently been 
found in the peripheral blood. In the blood-vessels invaded by sarcoma 
thrombosis is not as common as in carcinoma. Borst states that they 
have a certain capacity to grow in the lumen without connection with the 
vessel-wall. These considerations may tend to explain the frequency of 
metastases by blood-vessels. The metastases of sarcoma often occur much 
earlier than is generally conceived and this fact adds to the gravity of the 
prognosis after surgical operation. In a considerable series of cases at the 
Memorial Hospital the a:-ray photograph of the lungs has revealed pulmonary 
nodules in apparently operable cases, especially of bone-sarcoma, and in not 
a few which had just recovered from operation. 

Structure. — Sarcomas are chiefly liistioid tumors. With the exception 
of mixed or teratoid growths, a definite arrangement of elements in organoid 
form is missing. Yet considerable differentiation of tissues occurs in the 
more mature types. Fibrosarcoma may present areas resembling cellular 
connective tissue; bony trabeculae appear in many osteosarcomas; embryonal 
fat tissue occurs in liposarcoma; while angiosarcoma produces many func- 
tionating blood-vessels. 

With increasing anaplasia the form and function of the originating cells 
are gradually lost. 

The controlling factor in the structure of sarcomas is found in the natural 
tendencies of the cells of origin to reproduce the mother tissue. Many 
authors have traced in the growth of sarcoma a repetition of the embryonal 
development of the affected tissue. 

In many cases, the origin of the tumor is readily determined by the 
presence of cellular connective tissue, cartilage, bone or fat. 

In other cases the tumor is highly cellular and reveals few traces of the 
original form or function of its cells, and in the most anaplastic types the 
cells are round and completely undifferentiated. 

According to cell form, sarcomas may be divided into spindle-cell, 
round-cell and giant-cell groups, but these terms convey little information, 
and, unless the tissue or origin can be stated, no significant diagnosis has 
been reached. Since any one of the three above types of structure may 



arise in sarcomas from each mesoblastic tissue, it is desirable to avoid terms 
referring solely to the morphology of the cells. 

According to histogenesis sarcomas may be classified as fibroblastic, 
angiosarcoma, chondrosarcoma, osteosarcoma, liposarcoma, myosarcoma, 
myxosarcoma, lymphosarcoma, and gliosarcoma. In typical cases the 
structure of each of these forms is specific. 

In general the structure of sarcomas presents an imperfect development 
of the tissue of origin. The cells are usuUy larger and always more nurnerous. 
Very large spindle-cells belong chiefly to malignant fibroblastic, angioblastic, 
and myxosarcomas. Quite small spindle-cells also appear in tumors of these 
same origins. Giant-cells of various types appear in many sarcomas, 
especially in the tumors of bone and smooth muscle. In muscle they result 
chiefly from hypertrophy of tumor-cells; in bone-tumors they have been 
interpreted as giant osteoclasts; in bone-marrow tumors they are likened to 
the myeloplacques; in many cases they evidently form by cell fusion in 

Fig. 65. — Large spindle-cell fibroblastic sarcoma. 

poorly nourished areas; in the epulis they represent fused endothehum sur- 
rounding foreign material; while amitotic nuclear division is responsible 
for many. 

The origin and significance of the different forms of giant-cells is best 
considered in connection with specific tumors, since in many respects they 
are peculiar in each tumor and form an important means of diagnosis. 
Wakabayashi found the same groups of chromosomes and lack of mitoses in 
seven types of sarcoma as he observed in tuberculosis and syphilis. Howard, 
studying the morphology of many giant-cells, observed in the behavior of 
the nuclei many resemblances to protozoan nuclei. 

Round-cells appear in lymphosarcoma, gliosarcoma, and myelosarcoma, 
but it is doubtful if true round-cell tumors arise from any other mesoblastic 
tissues. While very cellular and anaplastic growths composed of indifferent 
spheroidal cells arise from many tissues, perfectly fixed portions of these 
tumors regularly show the cells to be spindle or polyhedral, and the effort 
should be made to detect these features so as to exclude the tumor from the 
group of true round-cell growths. 


Such distinctions are, however, rarely attempted and the term round-cell 
sarcoma is in very general use and enjoys recognition from practically all 
writers. Malherbe, however, separates sharply between round-cell sarcomas 
which he regards as exclusively lymphoid, and true fibroblastic tumors. 
Virchow stated that a round-cell tumor could safely be identified as sarcoma 
only when a definite intercellular stroma was demonstrated. Borst describes 
the cells of small round-cell sarcomas as identical with the round-cells infil- 
trating grariulation tissue, but the source of these cells he left undetermined. 
Highly mahgnant small round-cell tumors arise in the intermuscular septa, 
periosteum, submucous stroma, meninges, skin, kidney, testicte, and liver 
(Borst), and are said to represent the highest degree of sarcomatous degenera- 
tion (Rindfleisch). Yet it has never been shown and appears unhkely that 
any fixed connective-tissue cells in such numerous situations give origin to 
such tumors. 

From my own observations I have been forced to conclude that with rare 
exceptions such round-cell growths are either lymphosarcomas or small- 
cell carcinomas. They usually invade lymph-nodes. 

It is chiefly among the sarcomas of bone that small round-cell tumors 
appear which suggest an origin from periosteal fibroblasts, but here the 
proximity, of bone-marrow provides a possible origin from' lymphoid 

Much the same uncertainties surround the origin of most of the so-called 
large round-cell sarcomas. These tumors may, as a rule, be divided among 
endothehomas, tumors of lymphoid cells, carcinomas, and sarcomas of which 
the cells are not strictly round. 

The round-cell tumors of the testis and thyroid are among the epitheUal 
growths commonly misinterpreted as large round-cell and alveolar sarcoma. 

The nuclei of sarcoma-cells exhibit the widest variations in size, form, 
chromatin content, nucleolar elements, and methods of division; all of which 
features reflect rapid and lawless growth and the accompanying degenerative 
processes. All the abnormalities observed in carcinoma-cells are dupHcated 
or even exaggerated in sarcoma. In pronounced grade they are specific 
of a malignant tlimor process but not of sarcoma. Favre and Regaud 
find the same mitochondria granules iii sarcoma-cells as in carcinoma. 

Degenerative changes in the cells of sarcoma occur, chiefly in areas of 
necrosis from infarction. In actively growing, well-nourished zones the 
cell cytoplasm is usually less subject to fatty and hydropic degeneration 
than in carcinoma. Yet under appropriate conditions the cells undergo 
the usual forms of cytoplasmic decay. The most common change is a 
mucoid transformation of cells and stroma which affects nearly all forms of 
sarcoma, especially those arising from connective tissue. Extensive deposits 
of blood-pigment may lead to confusion with melanoma. Calcific deposits 
are comparativelj' rare, except in osteogenic tumors. Glycogen is found 
in the cells of many rapidly growing sarcomas (Brault) . 

The arrangement of the cells reflects the structure of the mother tissue 
and often gives a clue to the origin, yet in cellular growths all traces of the 
originating tissue may be lost. Prominent in many sarcomas is the tendency 
of the cells to grow about blood-vessels simulating the type of angiosarcoma 
or perithelioma. The specific structure of these latter tumors may be all 
but duplicated by tumors which have no angioblastic relations whatever. 
On the growing edges of infiltrating sarcomas the cells are commonly gathered 
about the small blood-vessels. A true alveolar arrangement is rarely if 
ever assumed by sarcomas. While it cannot be asserted that the small 
spheroidal or polyhedral sarcoma-cells never appear in groups surrounded by 


stroma, the history of most alveolar sarcomas has terminated in their identi- 
fication as carcinomas or endotheliomas. 

The stroma of sarcomas is derived from remnants of preexisting tissue, 
from blood-vessels appropriated for the nutrition of the tumor or new formed 
as an integral part of the neoplasm, and from the specific intercellular sub- 
stance derived from the tumor-cells. In relatively adult growths an inter- 
cellular stroma is regularly laid down by the cells, and in tumors of bone, 
cartilage, mucoid tissue and gha-tissue ■ it may form an easily recognized 
element. Yet in \-ery cellular tumors the attempt to demonstrate a stroma 
commonly fails and it would appear that the diagnostic significance of this 
element in sarcomas has been overstated. W. C. White found a fine inter- 
cellular reticulum very constant in sarcomas but absent in carcinomas. He 
identifies the reticulum of lymphosarcoma with white fibrous tissue. Seelig, 
in specimens digested by pancreatin, found a rather abundant fine reticulum 
in many sarcomas but failed to note any constant differences in the reticulum 
of lymphosarcoma from that of other round-cell sarcomas. Bielschowsky's 
stain is particularly effective in demonstrating a reticulum in sarcoma 
as well as in carcinoma. 

The blood-vessels form the most important part of the stroma. They 
are of various forms and sources. Following the analogy with, granulation 
tissue the blood-vessels may represent elongations of those of the originating 
tissue about which the tumor-Cells grow. Such vessels may be venous, 
arterial or capillary, and their walls are composed of adult normal cells. In 
the course of their growth they may become varicose or sinusoidal and the 
walls thin and fragile. Some of the well-formed vessels are probably appro- 
priated directly from the invaded tissue. As the tumor develops, the 
adventitia of the vessels is lost in the stroma of the neoplasm. Eventually, 
nothing but a swollen endothelial cell separates blood-current from tumor- 
tissue. Malherbe describes the apparent sarcomatous transformation of the 
muscle-cells of vessel-walls and points out the diagnostic importance of the 
swollen condition of the cells in the vessel-walls of sarcoma. Many blood- 
spaces and channels appear to be formed by modified tumor-cells but such 
structures are extremely fragile and leave the tumor-tissue subject to inter- 
stitial hemorrhage. 

In a large group of angiosarcomas and of peritheliomas the tumor-cells 
exhibit true angioblastic properties and the tumors are composed essentially 
of a congeries of very cellular blood-vessels. 

The continued growth of most sarcomas is closely dependent upon the 
capacity of the cells to excite the development of blood-vessels. 

Where the vascular channels fail, growth ceases, and when they become 
occluded anemic necrosis promptly follows. Yet some very cellular actively 
growing small spindle-cell sarcomas contain few vessels. 

Lymph circulation in sarcomas is very deficient and according to Ziegler 
lymph-vessels cannot be demonstrated. Yet in some varieties clefts and 
canals appear which probably represent lymph pockets or channels, and some 
cutaneous sarcomas have abundant lymph-vessels (Unna). Nerve- trunks 
invaded by sarcoma undergo pressure-atrophy. 

In neurosarcomas fragments of myelin sheaths and axis-cylinders may 
long persist. Definite new formation of nerve-fibers in sarcoma has not been 

Etiology. — Of the specific etiology of sarcoma little is definitely known. 
These tumors demonstrate the very great proliferative capacity of meso- 
blastic cells released from the restraints to growth. This capacity may 
reasonably be estimated as even greater than with most epithelial tissues. 


It is commonly assumed that normal adult cells are incapable of such 
great proliferation as occurs in sarcoma, and the further assumption is 
then necessitated that sarcomas as a rule grow from isolated, or superfluous, 
or embryonal cell groups. Thus when trauma is followed by an extremely 
malignant bone-sarcoma, the violent explosion of growth capacity seems to 
require some element which is not attributable to normal cells regenerating 
after contusion. The rarity of any definite observations of the beginnings 
of such sarcomas is a strong defense of this hypothesis. Moreover there are 
many instances in which sarcomas are known to arise from isolated or em- 
bryonal cell groups, especially in the mixed tumors of teratoid type. 

On the other hand it may be urged that the regenerative capacities of 
mesoblastic cells are very great, especially in youth when most sarcomas 
occur. It must be admitted that the existence of isolated cell groups as the 
source of many sarcomas has not been proven. 

Sarcoma has been observed to follow x-ray injury in rats, and in mice 
the continued growth of carcinoma is said to have excited sarcomatous pro- 
liferation of the tumor-stroma. Finally, increasing observations of very 
early sarcomas and of presarcomatous lesions of inflammatory origin favor 
the view that sarcomas often arise from previously normal adult cells. 

It is clear that no general rule can be applied in this field and that the 
evidence must be worked out separately in each form of sarcoma. 

Presarcomatous Lesions. — The occurrence of atypical productive 
inflammatory lesions leading to sarcoma must be regarded as fully in accord 
with established views regarding the nature of sarcoma. Many sarcomas 
show such marked histological resemblances to inflammatory processes 
that pathologists have long been inclined to accept in a certain sense the 
inflammatory or even the parasitic origin of certain sarcomas. The lympho- 
sarcomas especially stand in this position and Borst, among others, antici- 
pates the ultimate identificaion of this group of neoplasms with the infectious 
granulomas. In such an event the sarcoma must be regarded as an indirect 
result of the parasite, much as some carcinomas are an indirect sequel of 
chronic irritation. Syphilis and tuberculosis are the most prominent infec- 
tions which lead to sarcoma. 

Relation of Tuberculosis and Syphilis to Lymphosarcoma. — Clinical observa- 
tions have long indicated that tuberculous lymphadenitis may pass rapidly 
or after several recurrences into lymphosarcoma. Ricker in a case termin- 
ating after 14 years found typical tuberculosis associated with lymphosarcoma 
of neck, lungs, adrenals and spinal cord. Miiller saw general sarcomatosis 
from a primary tumor of the breast, miliary tubercles being nearly coexten- 
sive with widespread sarcomatous tumors. The rather frequent discovery 
of tubercle bacilli in lesions supposed to be pure lymphosarcoma suggests 
that a very shght difference separates some of such tumors from the immedi- 
ate or distant presence of tubercle bacilli or their toxins. I have followed 
a case of cervical tumors originally tuberculous, recurring four times in six 
years and terminating with the picture of endothehoma of lymph-nodes. 
In this study it was shown that an infectious granuloma may cause prohfera- 
tion of the reticulum cells, producing large-ceU lymphosarcoma, or of the 
sinus endothelium producing endothehoma of lymph-nodes, and it was 
also shown that exactly similar tumors of lymph-nodes occur in the axilla 
without any histological evidence of granuloma. Various diphtheroid bac- 
teria may be isolated from such tumors. 

The transformation of Hodgkin's granuloma into sarcoma has been 
described by various authors, but I have found evidence that most of these 




tumors arise from the thymus where again they are implicated with 

The difficulty of distinguishing from sarcoma certain late syphilitic 
lesions in bone and muscle is mentioned by many authors, especially by 
V. Hansemann. The chnical diagnosis of syphilis is rendered difficult by 
the recurrence of the lesions after operation, by their multiplicity, and by 
the resistance to syphilitic treatment, while the structure presents very 
marked cellular overgrowth of round-cells or spindle-ceUs. Yet the syph- 
ilitic process tends toward necrosis and cicatrization rather than progressive 
growth. There are many indications that true sarcoma develops from such 
lesions, but satisfactory proof is lacking. 

Fig. 66. — Structure of rapidly forming callus developing in muscle-tissue two weeks'after 
fracture of clavicle. Note resemblance to osteogenic sarcoma. 

In the healing of fractures and of wounds of the periosteum the histolog- 
ical picture shows very active growth of bone, giant-cells and connective- 
cells, and regenerating capillaries, and may distinctly resemble sarcoma. 
The actual transition into sarcoma has not been traced. 

Traumatic myositis ossificans appears to occupy a position intermediate 
between inflammatory and neoplastic processes (Berndt). The clinical 
history of established cases is that of a self-limiting productive inflamma- 
tion, but the histology of the early lesions may be diflacult to distinguish 
from bone-sarcoma. It appears that the division of cases is accidental, 
some of the lesions progressing as myositis, others going on as traumatic 
sarcoma. I have observed the two conditions combined. Central giant-cell 


sarcoma of bone has often been interpreted as a form of productive inflamma- 
tion, but its various phases show every gradation from the benign to a 
mahgnant process. Kohsko finds that osteitis fibrosa may be followed 
by progressive changes leading to giant-cell sarcoma of moderate maUgnancy. 

Many clinical observations point to the development of sarcoma from 
granulation tissue, and the histological study of granulation tissue and of 
organizing blood-clots occasionally reveals pictures which closely approach 
the structure of sarcoma. Heukelom has traced in considerable detail the 
gradual transformation of the cells of granulation tissue into sarcoma- 
cells. Occasionally one finds small tumors following trauma and containing 
much blood-pigment, of which the structure is distinctly sarcomatous. 

Against the hypothesis of the origin of sarcoma from previously normal 
cells stands the opinion of Birch-Hirschfeld, Borst and others that sarcomas 
do not develop from previously normal cells but from embryonal cell groups. 
This view must be regarded as an hypothesis without adequate proof. 
Ackermann was unable to find any evidence of the origin of a series of sar- 
comas from embryonal cells. 

It seems highly probable that sarcomas, hke carcinomas, arise through 
exaggerated inflammatory and regenerative overgrowth of tissue-cells. 
While in many cases the cells of origin are adult, theoretical considerations 
suggest that in some instances they are embryonal. 

In the former group only would specific presarcomatous stages be 

The clinical transformation from benign into malignant mesoblastic 
tumors has repeatedly been observed, but the exact nature of the changes is 
somewhat uncertain. There is no doubt that in a small proportion of cases 
of fibroma, uterine myoma, neurofibroma, and lymphangioma, a long stand- 
ing and quiescent tumor eventually displays rapid growth and the structure 
proves highly cellular and malignant. In rare instances there is every 
indication that the slowly growing cells have taken on greatly increased 
powers of growth. In other cases it has been assumed that the malignant 
process arises de novo in some element of the tumors, as the blood-vessels. 
For the assumption that the sarcoma develops from embryonal cell groups 
included in the benign tumors there seems to be very insufficient basis. 


Spindle-cell Sarcoma; Fibroblastic Sarcoma. — This common tumor 
represents the purest form of fibroblastic neoplasm. It is of widespread 
occurrence, developing in nearly all situations where supporting connective 
tissue is found. The chief locations are in the subcutaneous and submucous 
tissues, the fasciae and muscles, periosteum, and in the parenchymatous 

In the gross these tumors are usually single, but occasionally multiple. 
The form is rounded or lobulated and the borders are not sharply marked. 
The consistence is hard in the fibrous and small cell-growths, softer and more 
elastic with the large cell-tumors. The hard tumors are opaque and non- 
vascular, the soft ones are vascular, succulent, reddish and variously altered 
by secondary changes. The minute markings are as a rule not distinctive. 
In size one encounters very small nodules in the earliest stages and bulky 
fungating and necrosing growths in the later periods. In the skin they form 
multiple, nodular or globular growths which may become pol3^oid. 

The rate of growth varying with the structure may be very slow or very 
rapid. Cessation of growth is rare but cutaneous tumors may be sponta- 
neously extruded. Extremely rapid and bulky tumors are observed espe- 
cially after injuries. Being poorly circumscribed local recurrence frequently 
follows removal even with the comparatively benign forms. Metastases in 
lungs, liver, and other organs frequently occur with large cell-growths, and 
rarely, as in Walter's case, one is surprised to find secondary growths from 
apparently benign small spindle-cell sarcoma. 

The structure falls into two main classes which differ also in their clinical 
features: (i) small, and (2) large spindle-cell sarcoma. 

The small cell-tumors are slowly growing, relatively benign neoplasms 
which approach the structure of fibrosarcoma on the one hand and on the 
other may approximate the type of large cell- tumors. Their rate of growth, 
density and malignancy, usually accord with these distinctions. The large 
cell-tumors show more active growth, are softer, more vascular, readily 
suffer from necrosis, ulceration and hemorrhage, and often recur locally and 
produce metastases. 

The spindle-cells of either type resemble fusiform fibroblasts. They are 
much smaller than normal fibroblasts and more densely packed in the small 
cell-growths, but reach very large dimensions and are loosely arranged in 
the large cell-type. The cytoplasm is granular, opaque and acidophile, 
the nuclei vesicular and provided with one or more small nucleoli. 

In the compact tumors cell borders are distinguished with difficulty and 
nuclei occupy most of the field but when the texture is looser the cells exhibit 
elongated processes and an intercellular stroma becomes visible. This 
stroma when well developed is composed of fibroglia fibrils demonstrable 
by Van Gieson's or better by Mallory's stain. It may become abundant 
or be reduced to a trace or be altogether absent. Much of it radiates from 
the blood-vessels. Its diagnostic importance is not great. Besides the 
specific stroma of the tumor, remnants of invaded tissue often appear and 
these with the blood-vessels often determine the arrangement of the cells. 



The cells may run diffusely, but as a rule they lie in broad bundles or 
narrower columns following blood-vessels or inclosed in connective-tissue 
septa, or coursing in various directions so that longitudinal, cross, and oblique 
sections appear in the same field. Virchow and others have employed many 
terms to designate these structural details as sarcoma lamellosum, fascicida- 
tum, traheculare, radiare. As a rule these features do not give any clue to 
the histogenesis. 

The spindle-cell sarcoma tends to infiltrate surrounding tissues, the 
cells pushing their way between fat-cells, vessels, or gland structures, all 
of which suffer atrophy. The walls of vessel may be infiltrated, penetrated 
and destroyed, and since the vessels are an easy prey metastases by these 
channels are common (Goldmann, Hedinger). Bone is penetrated by 
way of the Haversian canals, and readily absorbed. 

The blood-vessels of spindle-cell sarcoma increase with the larger size 
of the cells and the more rapid growth. They present a variety of structural 
t}rpes. Scanty but well-formed vessels predominate in small cell-tumors 
but in others the vessels may be very numerous and the walls are composed 
of single layers of endothelial cells supported by little or no adventitia. In 
several forms the vascular paths consist of a system of sinuses walled only 
by tumor-cells. Thus hemorrhage is common and metastases by way of 
blood-vessels is facilitated. While in many cases the vessels appear not 
to belong to the neoplastic process, as they become more abundant the 
cells usually appear hypertrophic and their nuclei hyperchromatic. The 
indications of lymph-paths are usually wanting, but Unna describes abundant 
dilated lymph-spaces in certain cutaneous sarcomas. In defective lymph- 
paths V. Heukelom saw a factor in the pathogenesis of sarcoma, but Lubarsch 
properly interprets the absence of lymph-channels when it exists as a result 
and not the cause of the sarcomatous process. 

Secondary changes in spindle-cell sarcoma are not prominent. Fibrosis 
may appear in slowly growing small cell-tumors and an arrest of growth may 
rarely be established. Wide fibrous areas sometimes appear in large cell- 
tumors as a local result of occlusion of vessels. More often hemorrhage 
and necrosis follow rupture or thrombosis of vessels, and pseudomelanosis 
frequently recalls old extravasations of blood. Chronic edema affects old 
and large tumors in which venous stasis occurs, and a pseudomyxomatous 
change may overtake small foci or the whole tumor. Edema, fatty degenera- 
tion and simple necrosis combine to produce cysts with serous or blood- 
stained contents. 

In many soft vascular tumors the cells are polymorphous, small and large 
spindle, polyhedral and round-cells appearing in foci or throughout. Such 
growths are often called round-cell sarcoma but a true round-cell sarcoma of 
fibroblastic origin probably does not occur. Likewise polynuclear cells 
may form and reach considerable dimensions but the term giant-cell sarcoma 
is better limited to other classes of tumors. 

Types of nuclear division are extensively illustrated, especially in the 
large cell-tumors. They occur in the elongated spindle-cells and especially 
in large round-cells mingled with them. Increasing numbers of round-cells 
result from multiplication of dividing cells and constitute a sign of malig- 
nancy. Many of the anomalies of cell-division in tumors appear in the 
fibroblastic sarcomas (Stroebe). 

The diagnosis of spindle-cell sarcoma seldom presents difficulty but there 
are some notable exceptions to this rule and the proper erminology is still 
a matter of discussion. Typical structures occurring in ordinary situations as 
skin, bone, etc., are readily identified. When cell-fibrils become prominen 


and are of adult type the tumor is relatively benign and deserves the designa- 
tion fibrosarcoma. The numerous small cells and abundant fibrils serve 
to distinguish most fibrosarcomas from cellular fibromas which show more 
resemblance to large spindle-cell sarcoma. From myosarcoma the fibro- 
blastic tumors may be distinguished by their location, the large vesicular 
nuclei, and lack of acidophile quality of the cytoplasm. Yet the myogenous 
origin of certain spindle-cell sarcomas has often been suggested and is well 
worthy of consideration (testis, kidney). Very vascular tumors may be 
called telangiectatic or angiosarcoma, but the latter term has repeatedly 
been condemned as wrongly inferring an origin from blood-vessels. The 
malignant (sarcomatous) angiomas have quite a different structure and 
origin. From typical perithelioma the large spindle-cell sarcoma can some- 
times be distinguished with difficulty, but the perivascular units of radiating 
cells and the characteristic large giant-cells of the former growth are usually 

Large spindle-cell sarcoma may be closely simulated by various tumors 
whose existence must be specially noted. Malignant epitheUal tumors of 
the thyroid are very prone to assume the spindle-cell type, as do also the 
embryonal forms of the so-called hypernephroma. Metastatic melanoma- 
may appear as a spindle-cell or large round-cell unpigmented tumor. Not 
a few epidermoid carcinomas present spindle-cells throughout most of their 

Diffuse endothelioma may exactly reproduce the structure of fibro- 
blastic sarcoma, but the location of the former growths, the appearance of 
typical endothelial characters in small areas, the arrangement of the cells, 
and their minute structure usually but not always permit a positive separa- 
tion of these closely allied types. 

Histogenesis. — In the origin of spindle-cell sarcoma the adventitial cells 
of the blood-vessels appear to take a prominent part. The early stages of 
these tumors have many times been traced to foci of vascular proliferating 
connective tissue which has been identified on doubtful evidence with granu- 
lation tissue. Some of these tumors appear to follow trauma, and the 
processes which result from solution of continuity, extravasation and absorp- 
tion of blood, often bear considerable resemblance to sarcoma. In organizing 
blood-clots there may be very active growth of vessels about and into the 
clot and it may be supposed that an exaggerated growth of these structures 
leads in certain cases to sarcoma. Likewise in some established and sponta- 
neous tumors the cells appear to be derived from proliferating adventitial 
fibroblasts. Such appearances are rather frequent and occasionally con- 
vincing, and Babes and Ackermann have expressed the opinion that fibro- 
blastic sarcomas in general originate from blood-vessels. Lubar?ch also 
described spindle-cell sarcomas in young subjects which on close analysis 
reveal themselves as obliterating cellular angiomas. 

The part played by endothelial cells in the genesis of sarcoma has been 
variously estimated but in the pure spindle-cell tumors they are not promi- 
nent. Many endotheliomas, however, closely resemble this type of sarcoma 
and since it is desirable to adhere to the fibroblastic origin of spindle-cell 
sarcoma the endothelial tumors must be excluded from this group as long 
as endothelial characters are retained. 

Smooth muscle-tissue may also give origin to tumors of spindle-cells, 
as in myosarcoma uteri, but these tumors also must be separately classified. 
The intermuscular fasciae and the peri- and endomysium give rise to fibro- 
blastic sarcoma. Those occurring in the muscles of the thigh produce 
bulky tumors with a notable tendency toward mucoid degeneration. It 


appears probable that in certain situations {e.g., testis) spindle-cell sarcoma 
represents a one-sided development of teratoma, but to what extent this 
principle applies elsewhere remains to be determined. 

Sarcoma of the Skin. — The skin is the primary seat of a variety of sarcomas 
or sarcomatoid processes, the exact nature and position of which have not 
been determined. 

Some of these are of spontaneous origin, extend by displacement of 
normal tissues, grow progressively, produce metastases and cachexia, exhibit 
the definite histology of a malignant tumor usually composed of spindle- 
cells, and belong clearly in the group of sarcomas. Others, while producing 
single or multiple tumor-like growths, progress by infiltrating the tissues, 
may often regress or disappear, do not produce metastases although appearing 
in numerous foci, and they present in their earliest stages or throughout the 
histology of an inflammatory lesion which, however, may sometimes assume 
a more or less distinct neoplastic character. These conditions are often 
called sarcoids of the skin. 

Of the latter group many types have been described (Pollard, Lit.), 
of which some have been finally disposed of among infectious granulomas. 
Such are: (i) mycosis fungoides, types of which are often mistaken for sarcoid 
(Paltauf); (2) multiple benign sarcoid of Boeck, which proves to be a form 
of tuberculosis; (3) various cases of benign sarcoid which consist of diffuse 
lymphocytic infiltrations, and which, while sometimes reaching considerable 
dimensions, show a distinct tendency toward regression (Spiegler, Joseph, 
Pollard). Numerous other peculiar cases have failed to receive a final 

The chief general significance of the study of these cases lies in the evi- 
dence which they offer that various inflammatory conditions may closely 
approach the true sarcomatous process. Spontaneous regression is perhaps 
not a sufficient ground, as Fano claims, for the exclusion of such processes 
from the group of tumors, since other true sarcomas may occasionally regress, 
but the entire clinical and histological picture justifies the current view that 
these processes are pseudo-neoplasms. Not clearly separable from the 
regressing sarcoids, but distinguished from them by its capacity to develop 
genuine malignant tumor characters is the multiple hemorrhagic sarcoma of 

Multiple Hemorrhagic Sarcoma (Kaposi). — This interesting malady, 
first fuUy described by Kaposi in 1872, has attracted much interest from 
dermatologists. Recently the data have been fully analyzed by Mariani. 

Observed chiefly in southern Europe, in elderly subjects, it exhibits 
a definite relation to certain predisposing factors, as general or local vascular 
lesions, sclerosis, ectasias, hemorrhagic diathesis, edema, and to trauma and 
alcoholism. A prodromal stage is marked by local cyanosis, edema and 
slight -inflammation, or by definite vascular ectasiae (De Amicis). In the 
stage of tumor growth multiple soft bluish nodules or flat infiltrations form in 
the derma of the outer surface of hands, feet, cheeks, nose and elsewhere. 
They are often painful, inflamed, and may ulcerate, involving deeper tissues 
and even eroding bones. The nodules may be very numerous, Babes observ- 
ing 450 in one patient. Hemorrhage often occurs leaving pigmentation 
prominent in some cases. The disease progresses from the extremities 
upward, some areas showing a tendency to heal, especially under arsenic. 
The lesions may be symmetrical. In the stage of cachexia the cutaneous 
lesions may become numerous or diffuse and in several fatal cases secondary 
nodules appeared in the pharynx, gastro-intestinal tract and liver, kidney, 
pancreas, lungs and serous membranes (Phillipsohn,'"^Mariani, Lit.). Most 


of the cases are eventually fatal from hemorrhage, diarrhea, fever and 

The histological structure exhibits in the early stages chiefly new and 
dilated lymph- and blood-vessels and infiltration by lymphocytes, plasma- 
cells, mast-cells, and edema, and the lesion resembles that of an infectious 
granuloma. According to W. Pick the early lesion is indistinguishable 
from elephantiasis lymphangiectatica. Proliferation of endothelium and 
perivascular spindle- and round-cells may appear early and become so 
prominent as to lead to the diagnosis of angiosarcoma, etc. _ In the later 
stages spindle-cells may predominate and the lesion is indistinguishable 

Fig. 67. — Multiple hemorrhagic sarcoma. Kaposi. (After W. Pick, A.D., 87.) 

from spindle-cell fibroblastic sarcoma (Pick). In all stages vascular ectasise, 
hemorrhage and pigmentation are constant or frequent. 

Regarding the nature of this disease opinions are naturally at variance, 
but the conclusion seems well founded that we have here to do with an 
infectious granuloma of unknown origin, which in its later stages in pre- 
disposed subjects and under suitable conditions may take on genuine neo- 
plastic properties. Justus seems to have carried the agent through five 
generations of rabbits with the production of typical lesions. The secondary 
growths in the viscera are probably not metastases but extensions to new 
originating foci. 

Beginning in local lesions in the extremities in which an infectious agent 



is probably concerned and aided by the various predisposing factors, the 
entire vascular system seems to be overtaken by degeneration with over- 
growth of vasoformative cells leading in many instances to true tumor 
formation. In the histogenesis the numerous discussions have dealt with 
all the data regarding the interrelations of endotheUum, peritheUum, fibro- 
blasts, plasma-ceUs, and polyblasts, but the final tumor product is a spindle- 
cell sarcoma, which on account of its early history and tendency toward 
regression has often been called sarcoid. 

Spindle-cell Sarcoma of Skin. — These tumors appear as single large 
growths which tend to become pedunculated or as multiple small nodules 
widely distributed. Appearing in the derma or subcutaneous tissue they 
form circumscribed nodules or masses which vary widely in color .^consist- 

FlG. 68. — Multiple hemorrhagic sarcoma. Kaposi. Intestinal lesion. 

A.D., 98.) 

{After Mariana, 

ence, as well as in number and position. Some are hard, opaque and fibrous, 
others soft, reddish, vascular and edematous. Both the single and multiple 
tumors are usually of slow growth and relatively benign, but either type may 
develop greater malignancy and terminate with metastases, and cachexia 
(Unna). As a rule these tumors show a hmited capacity for growth and 
while recurring locally are not directly fatal. Fibrosis (Linser), peduncu- 
lation, and spontaneous extrusion may occur. Ulceration and necrosis are 
rare. They appear at almost any age and Maldovan and Fernet describe 
congenital cases. 

The structure is in many instances that of the simplest form of small 
or large spindle-cell tumor, but many variations from this type are encoun- 


tered. Blood-vessels may be numerous and the tumor-cells may appear 
to be derived from their adventitial cells. Varying in this direction spindle- 
cell sarcoma of the skin approaches angiosarcoma and perithehoma from 
which types they are not always sharply separable. In one group endothelial 
prohferation appears to be prominent and yields many rounded and poly- 
hedral cells (Johnston). Chronic edema may lead to secondary myxoma- 
tous characters. Unna describes a rich supply of lymph-spaces in some 

The arrangement of cells also varies greatly. They may be closely 
packed, of small size, with little stroma, and quite diffusely arranged. An 
interfascicular form is yielded by cells whose axis is determined by the course 
of blood-vessels (Unna). The cells may form whorls or intertwining bundles 


^ •''^: %& 

? X :. 'ji . '*i;, 


Fig. 69. — Structure of multiple hemorrhagic sarcoma. (After Mariani.) 

which have suggested an origin from nerve-trunks. The cells penetrate sur- 
rounding tissues and develop new outlying nodules, or they may remain well 
circumscribed. In the central portion of all true sarcomas of the skin Unna 
finds elastic fibrils missing. 

The histogenesis of cutaneous spindle-cell sarcoma is a complex question, 
and it has not been possible to clearly define the subgroups probably exist- 
ing in this class according to their exact origin. Yet the fibroblastic nature 
of the typical form seems reasonably certain. Many have concluded that 
the adventitial cells of blood-vessels and lymph- vessels are the chief seats of 
origin. In certain cases I have been impressed with the evident relation 
of the tumor-cells to numerous abortive blood-vessels which seemed to be- 
come lost in the tumor-tissue. Yet Unna believes that the importance of 
blood-vessels in the origin of these growths has been overestimated, espe- 



cially by Babes. The connective tissue about cutaneous glands rarely seems specially concerned. That an endotheUal element is present in some 
cases is highly probable, especially in tumors with rounded, polyhedral or 
giant-cells. In cases of multiple symmetrical tumors with cells in inter- 
twining bundles an origin from nerve-sheaths and an endothelial element are 
indicated. The occurrence of congenital cases indicates that misplaced or 
superfluous cell groups, and not normal tissues, are occasional or frequent 
points of origin. Finally it is clear that some tumors closely resembling 
fibroblastic sarcoma belong in the group of nevomelanomas, and that the 
highly malignant cutaneous sarcomas as in Pini's case, must be regarded with 
suspicion from this standpoint. 

In the etiology, various forms of trauma not infrequently precede the 
sohtary sarcoma (Lowenstein) . Virchow (II, 245) cites many cases of 

Fig. 70. — A giant-cell tumor of tendon sheath, probably of endothelial origin. 

sarcoma arising in scar tissue which had been subjected to long irritation. 
For the multiple tumors a tissue predisposition and the various inflammatory 
and nutritional changes illustrated in the sarcoids may be held responsible. 

Giant-cell Sarcoma of Tendon-sheaths and Aponeuroses. — The specific 
structure and benign clinical course of a group of giant-cell sarcomas of 
tendon sheaths of the hand and feet have long been recognized. They 
have been fully described, especially by French observers under the terms 
myeloma or xanthosarcoma (Gross, Paquet, Reverdin, Heurteaux, Spiess, 

The tumors arise chiefly in tendon sheaths and aponeuroses of hands and 
feet, and reach moderate dimensions after a period of i to 20 years. Sudden 
increase in size may rouse a suspicion of malignancy, but they seldom reach 
the size of an egg. They are well encapsulated, firm, opaque or j'-ellowish 
and markedly lobulated, and are readily shefled out by operation. They 
never yield metastases and rarely recur after removal. 


The structure presents four main features, (i) Groups of spindle- or 
polyhedral cells surrounded by dense connective tissue. Some of the 
tumors are exclusively of this type. Although the tumor may be quite 
cellular and free from giant-cells it still maintains its benign character. (2) 
Groups of large cells containing much doubly refractive lipoid material. 
These are the so-called xanthoma-cells. They may be so abundant as to 

Fig. 71. — Myxosarcoma of fascia of thigh. 

duplicate the appearance of xanthoma of the skin, and their origin is referred 
to proliferating fibroblasts or endothelium. (3) Giant-cells of very large 
size up to 125 m. are commonly present in abundance, and present very 
numerous small vesicular nuclei. The cytoplasm is granular, acidophile, 
and often contains much lipoid material. They belong to the class of foreign 
body giant-cells. (4) Pigment of the type of hemosiderin frequently collects 
in the tumor-cells. 

Etiological factors are very indefinite but the structure and benign char- 


acter of the process suggests that the tumors arise from a local inflamma- 
tory process connected with trauma and some peculiar disturbance of local 
lipoid metabolism. Fleissig denies that they have any neoplastic quality 
and regards them as a form of granulation tissue, yet similar arguments 
apply almost equally well to many other sarcomas. They have often been 
classed with xanthoma. 

Sarcoma of Muscles. — Primary sarcoma of the muscles arises from the 
connective tissue and vessels of the perimysium and endomysium and usually 
takes the form of fibrosarcoma or myxosarcoma with small or large spindle- 
cells, and often with a marked prominence of capillary or larger blood- 
vessels. Occasionally they are described as round-cell sarcoma (Chambe, 
Pasteau, Kuttner). 

They occur chiefly in young adults, the rapidity of growth varies with 
the structure but is often rapid, they reach a large size if undisturbed, recur 
locally after operation, and the cellular forms produce metastases. Chambe 
has collected 39 cases illustrating many features. Many cases were fatal. 
Virchow observed that -in the muscle fasciae slowly growing fibrosarcoma or 
spindle-cell sarcoma predominate, while at tendon insertions many round- 
cell and malignant tumors appear. A characteristic fibrosarcoma arises 
from the sheath of the rectus abdominis (Bodenstein, Lit.). 

The histogenesis has been traced in certain cases to the interstitial tissue 
(Birch-Hirschfeld), and the remarkable proUferative tendency of this 
tissue renders such an origin probable. The muscle-cells usually atrophy, 
but Sokalow, Guitton and others reported the transformation of muscle- 
cells into sarcoma-cells. About many tumors of muscle there is a pronounced 
reactive myositis with various alterations in the form of the cells (W. 

That sarcoma of muscle may be derived from groups of embryonal and 
undifferentiated muscle-cells has often been suggested (Borst). It seems 
probable that with myxosarcomas and malignant growths approaching the 
round-cell type the origin is from misplaced cell groups connected with bone 
formation. In a series of bulky sarcomas of the thigh I have observed 
gradations in structure from myxosarcoma to chondrosarcoma. In an early 
sarcoma of muscle, I found very many small blood-vessels which seemed to 
be the source of the spindle-cells of the tumors. Peabody's study of Peters' 
cases suggested a vascular origin. Sokalow traced some tumors to the sarco- 
lemma, but Cornil and Ranvier, Malherbe, and most authors assume that 
they arise from the interfibriUar connective tissue. 

Periosteal sarcomas of many types commonly invade the muscles and 
some may appear largely intramuscular. 

Sarcoma of Esophagus. — Sarcoma of the esophagus is a rare condition. 
Howard (1902) analyzed 12 cases. Herxheimer, 1908, refers to 18 cases, 
two of which were combined carcinoma and sarcoma. Hacker, 1909, gave 
a complete analysis of the subject with reference to 21 cases. They occurred 
usually in males, in the upper or lower segment, and in subjects from 4_ to 
70 years of age, and were rapidly fatal. They appear either as an ulcerative 
process or as a polypoid tumor. The former usually contain round- or poly- 
hedral cells, the latter, spindle-cells. Stephan describes a pecuUar lympho- 
sarcoma in an infant. In Eucken's case there were many giant-cells. With 
the round-cell growths metastases are early and widespread (Shaw, Stark, 
Rolleston). The spindle-cell tumors were usually free from metastases 
(Targett, Ogle, Livingood, Herxheimer). 

It seems probable that two entirely different tumors have been included 
in this group. The round-cell growths strongly resemble embryonal card- 


noma in gross appearance, and Stark suspected that his two cases might be 
carcinomatous. One case of this type with bulky metastases in the Hver 
strongly suggested to me a carcinomatous origin (Norris). The spindle- 
cell tumors are probably genuine sarcomas, but little effort has been made 
to determine their origin. Howard regarded his spindle-cell -tumor as a 
myosarcoma. Wolfensberger and Glinsky have observed mixed tumors of 
the esophagus composed of striated muscle-, spindle-, and giant-cells. Herx- 
heimer describes combined sarcoma and epithelioma arising apparently 
from the submucosa and the epithelial lining. Metastases were wanting. 
Baur describes a melanotic sarcoma of esophagus. 

Sarcoma of Stomach. — The reported cases of sarcoma of the stomach 
constitute about i per cent, of all gastric tumors. On analysis the cases fall 
into three distinct groups: (i) Spindle-cell myosarcoma; (2) lymphosarcoma, 
and (3) miscellaneous round-cell or alveolar sarcomas, the nature of which is 
uncertain. Of all these 150 cases have been collected by Ziesche and 

I. The spindle-cell myosarcomas constitute a well-defined group. They 
form solid or cystic growths, single or multiple, which project from the wall 
of the stomach into the peritoneum, or remain interstitial, or protrude into 
the cavity of the organ where they suffer necrosis and ulceration. Very 
early small growths have been observed, while Brodowski's tumor grew to the 
size of a child's head between the mesenteric layers. Cantwell's case weighed 
12 pounds and Baldy's nearly filled the abdomen. Most of these tumors 
arise from the curvatures, rarely from the pylorus (Manges, Schlesinger). 

The external pedunculated tumors may draw the stomach out into a 
funnel form; the interstitial tumors may cause dilatation; and Ewald found 
the organ reduced to a narrow canal surrounded by multiple or diffuse masses ' 
of fibrosarcoma. Ulceration and hemorrhage frequently occur. The 
large tumors are often cystic, and the bulky metastases which commonly 
form in the liver are often cystic and may transform this organ into a series 
of large cystic tumors (Hosch, Lit.). The course of gastric myosarcoma 
is relatively slow in the typical cases (three and one-half years) but with 
greater anaplasia of the cells the growth is often quite rapid. 

The structure varies between considerable limits. In some cases the 
type approaches that of a cellular myoma and the recognition of a myogenous 
origin is readily accomplished. Or the cells are of large spindle-form and 
the designation as a myosarcoma depends on the attitude of the observer. 
Such a case is that of Hoesch-Kaufmann which appears to be identical in 
structure with others described as large spindle-cell sarcoma. There is 
little doubt also that some or many of the so-called large round-cell sarcomas 
are of muscular origin. In some cases careful examination has revealed 
typical myosarcoma in one portion of the tumor, indifferent spindle-ceUs 
in other portions, and elsewhere only round-cells (Moser). The arrange- 
ment of the cells in intertwining bundles is often a characteristic feature. 
The grouping of cells about blood-vessels has sometimes suggested that the 
tumor arose from the walls of the vessels as in myoma uteri (Kaufmann, 
Howard), or it has led to the diagnosis of angiosarcoma, or lymphangiosar- 
coma (Steudel). The gross appearance of many of these tumors of doubtful 
nature is very similar to that of definite myosarcoma. 

In the etiology of myosarcoma a congenital anomaly must be sought 
for the stomach as for the uterus. Of the nature of this embryogenic dis- 
turbance there are scant indications. That isolated islands of muscular 
tissue occur in the stomach is shown by Alsleben's report of an adenomyoma 


in the pyloric mucosa, while the congenital occurrence of myosarcoma is 
probably demonstrated by Finlayson's case. 

2. Lymphomatous tumors of the stomach occur (i) as a part of the leu- 
kemic process; (2) in gastro-intestinal pseudoleukemia; (3) as a part of gen- 
eral lymphosarcoma, and (4) in the form of localized lymphosarcoma. They 

■ are considered in detail under these headings. 

3. Miscellaneous Sarcomas of Stomach. — In not a few gastric sarcomas 
the structure has appeared not to fall into any of common groups and the 
tumor has been variously described as alveolar or mixed-cell, or angiosarcoma, 
or myxosarcoma. None of these histological features deserves special 
recognition as a separate variety of gastric tumor. 

Dreyer described a pyloric tumor composed of spindle-cells and carcinoma 
occupying a large ulcerated area at the pylorus. The metastases were 
carcinomatous. Leube also observed mixed carcinoma and sarcoma. Such 
observations suggest the conclusion that carcinoma of the stomach may 
give histological appearances resembling sarcoma. Cammidge has men- 
tioned four cases of this type and concludes that many so-called sarcomas 
of the stomach are really carcinomas (Wilson) . This criticism appears to be 
justified and the writer would urge great caution in interpreting as sarcoma 
any tumor associated with an ulcer and of which the gross appearance 
strongly suggests carcinoma. That sarcoma of the stomach may follow 
ulcer appears possible (Moser, Kehr), but this relation has not been satis- 
factorily proven. Eighteen cases of metastatic sarcoma of the stomach, 
chiefly lymphoid and melanotic, have been collected by Ziesche and 
Davidsohn. They are not uncommon. 

Sarcoma of Intestine. — The small and large intestines and the rectum 
are subject to the same types of sarcoma as the stomach. Baltzer called 
attention to the surgical aspects of this group of tumors in a report of 14 cases. 
Rademacher (1908) collected 140 cases. Of rectal sarcomas Key found 
55 reports to which he added three. The majority of these cases were 
studied exclusively from the clinical standpoint, which emphasizes the 
symptoms of an abdominal tumor causing intestinal obstruction. The 
nature of the growths is seldom clearly stated but the great majority appear 
to be lymphomatous, while a considerable proportion are fibroblastic. 

1. Spindle-cell sarcoma arises chiefly from the muscular coat, and pro- 
duces a large, solid or cystic growth lying external to an intact mucosa. A 
rather well-defined group is the myxosarcoma affecting chiefly the cecum 
and appendix and said to arise from the submucosa. It may contain large 
cysts. The subserous tissue is said to give origin to certain angiosarcomas. 
A congenital angiosarcoma of ileum is described by Stern, and a subserous 
angiosarcoma of the splenic flexure by Baltzer. I have examined a large 
spindle-cell vascular sarcoma of the omentum. 

2. Miscellaneous sarcomas of intestine constitute a small group of ill- 
defined and imperfectly described varieties. An interesting case is that of 
Lehmann who designated as interfascicular endothelioma a diffuse thicken- 
ing of the entire intestine with secondary growths in liver. Pigmented 
sarcomas occur in the rectum and ileum (Sandner, Treves), but the source 
of the pigment is not clear and intestinal metastases of melanoma of the 
skin are common. I have studied one slightly pigmented primary sub- 
mucous melanoma arising around the opening of the appendix in a child 
(cf. Melanoma). 

3. Lymphomatous tumors of the intestine are considered under that 
heading. They arise chiefly in ileum, colon and rectum. Rarely they 
develop in the appendix (Warren, Carwardine) . 


Sarcoma of Ovary. — Primary sarcoma is a comparatively rare tumor 
constituting about 5 per cent, of all ovarian neoplasms (Kroemer). Yet in 
this small group many histological varieties are represented, concerning the 
histogenesis of which little is known. The process of organization of blood- 
clot in the corpus luteum and of regression of this structure offers conditions 
very favorable to the development of sarcoma, and it is probable that this 
source of myxosarcoma, perithelioma, etc., has not been given due atten- 
tion. Other sarcomas, especially of the embryonal type have been traced to 
a teratoid origin. 

1. Spindle-cell sarcoma is a well-defined type which is closely related 
to fibroma and probably also to ovarian myoma. It appears as a diffuse 
process affecting the whole ovary and producing either a solid tumor of 
moderate dimensions or a large growth containing many cysts. The surface 
is nodular or lobulated and section reveals a smooth opaque texture, or many 
dilated vessels, or cysts which are more irregular in size and form than in 
cystic adenoma. Areas of edema and mucoid degeneration may be present 
and hemorrhage or infarction may follow torsion. 

The structure presents spindle-cells of moderate dimensions mingled 
usually with spheroidal cells. Trabecule of adult connective tissue separate 
the tumor into lobules, while between the cells there is little demonstrable 
stroma. The cells may grow diffusely but more often they are definitely 
connected with small blood-vessels which are either thin-walled and telan- 
giectatic or compressed or occluded by hyaline degeneration. The main 
feature of some tumors is the abundance of hyaline vessels between which lie 
scanty spindle-cells. In the cystic tumors dilatation of lymph-vessels, 
edema and liquefaction necrosis, seem responsible for the irregular cavities 
which contain serous or bloody fluid and are lined by a ragged layer of sar- 
comatous tissue supported by connective tissue. Such growths may be 
interpreted as cystic lymphangiosarcomas. The spindle-cell tumors are 
of slow growth and relatively benign. 

The origin of the spindle-cell tumors must be referred to the ovarian 
stroma but to which elements is uncertain. The perivascular arrangement 
of the cells suggests an origin from the walls of vessels but not an endotheUal 
character. The perithelial tissue has been suggested as a possible source 
but without definite basis, since the tumors are usually diffuse. Smooth 
muscle-cells have been observed in round-cell and spindle-cell sarcoma 
and it appears probable that some of these tumors are of myogenic origin 
(Basso, Gangolphe, Kroemer). 

2. Round-cell sarcoma of the ovary is an ill-defined and rare form of 
malignant tumor which occurs chiefly in young subjects or children and 
proves rapidly fatal. The tumors are single or bilateral, reach a considerable 
size, are usually solid or contain softened or necrotic areas. They are soft, 
edematous, infiltrated with blood, and exhibit numerous necrotic foci. The 
structure is of small round-cells lying diffusely or grouped about thin vessels. 
In some cases the cells are of larger size and exhibit an alveolar arrangement 
(alveolar sarcoma). Or spindle-cells, small round-cells, large spheroidal 
cells and giant-cells may be variously combined. 

In none of these cases does the structure offer any clue to the histogenesis 
and this group of tumors therefore invites subdivision according to the 
histogenetic principle. It seems probable that some of the tumors with in- 
different round-cells are to be interpreted as one-sided developments of 
teratoma, but to which forms this view may apply it is difficult to decide. 
One form of large round-cell sarcoma of the ovary closely resembles a similar 


tumor of the testis and is probably of teratomatous nature. It is very 
prone to necrosis. 

It also seems probable that some so-called alveolar sarcomas are really 
of epithelial origin arising from the cells of the granular layer of the follicles, 
since these cells in adenomatous tumors readily assume a small round form 
and diffuse arrangement. That the theca cells of the ovarian follicles give 
rise to certain sarcomas is claimed by Pinto but this origin still lacks definite 

The relationship of cells to blood- and lymph-vessels has often led to the 
designation of such tumors as angiosarcoma and perithelioma. In favor of 
this view is the frequency of angiomas in the ovary and of -sarcomas in which 
the blood-vessel is the prime unit, yet most of these tumors are composed of 

Against the origin from vessels is the well-known tendency of all tumor- 
cells to thrive best in the proximity to blood-vessels, giving spurious angio- 
sarcoma. Nevertheless Kroemer recognizes a special group of ovarian 
angiosarcoma of blood and lymph-vascular origin. They are difficult to 
distinguish from the group of perithelioma (q.v.). In this field also one 
encounters evidences of the close relationship of endothelioma and sarcoma. 
The existence of a true ovarian endothelioma is elsewhere discussed. 

A form of plexiform sarcoma is, described by Kroemer. It consists of 
anastomosing columns of syncytial and giant-cells which he derives from the 
perivascular cells about lymph-vessels, a lymphangiosarcoma. The endothe- 
lial cells of the vessels proliferated slightly in places but tended to disappear 
from the field. This nice histological analysis still leaves untouched more 
important questions of histogenesis. 

Finally among the polymorphous cell ovarian sarcomas there remains 
a rather well-defined type. 

Myxosarcoma. — Areas of secondary myxomatous change are frequently 
seen in other forms of ovarian sarcoma. They occur about the blood-vessels 
and lymph-vessels, especially of fibrosarcoma and sometimes become so 
prominent as to call for the designation of myxosarcoma. Yet such cases 
are rare. Segalowitz found only one largely myxomatous tumor among 300 
from the ovary, and Kroemer states that a pure myxosarcoma of the ovary 
does not exist. An interesting observation is that of Walker who records 
three fatal myxosarcomas in sisters. 

Myxomatous areas occur especially with chondrosarcoma as in the case 
of Gibb which was probably of teratoid origin, and myxomatous changes 
in the stroma of ovaries invaded by epithehal tumors is so common as to 
lead Glockner to regard it as a specific reaction of ovarian tissue. A specific 
variety of myxomatous tumor of the ovary has been described by Krunken- 
burg with Marchand's endorsement, under the term fibrosarcoma mucocellu- 
lare carcinomatodes (see section. Carcinoma of Ovary). 

Chondrosarcoma occurs in the ovary as a one-sided development of a 
mixed tumor or of a teratoma, or as a metaplastic growth derived from the 
ovarian stroma. 

The former class is illustrated by tumors in which islands of cartilage 
are associated with dermoids, or with other derivatives of the ectoderm or 
entoderm. The cartilaginous elements may predominate and a character- 
istic case of this sort is that of Reis in which a large chondromatous tumor 
after extirpation was followed by a rapidly growing carcinoma. A similar 
origin probably applies to Gibb's myxochondrosarcoma occurring in a child 
of two and one-half years. Donati observed very large giant-cells in a 
tumor containing cartilage and much round- and spindle-cell tissue. Here 



it seems impossible to decide wiiether the origin was teratomatous or 

Some authors assume that ovarian and other chondromas of the female 
genital organs arise from inclusion of the Mullerian or Woffian ducts (Wilms), 
and Kehrer states that such tumors contain areas of embryonal round-cell 
mesodermal tissue which differentiate into various tissues. 

That cartilage arises by metaplasia from the ovarian stroma was strongly 
suggested by the old observation of Kiwisch who found cartilaginous plates 
in the outer layers of a bilateral ovarian fibroma. Buet also described 
cartilaginous metaplasia in a fibrochondroosteoma of the ovary. These 
processes were diffuse involving the entire ovary. The most notable evidence 
of metaplastic development of cartilage in the ovary is furnished by the case 
of Jung. 

In a multipara of 48 years with cervical endothelioma (epithelioma?) a myomatous 
uterus was removed with one-half of one ovary which was very slightly enlarged. This 
ovary was found to contain several islands of young and some ossifying cartilage embedded 
in edematous connective or in spindle-cell sarcomatoid tissue. There was a prompt recur- 
rence of a pelvic tumor at the site of the ovary containing chiefly chondrosarcoma with 
areas of spindle, round, and large polyhedral cells. Meyer interprets this case as a teratoma, 
Kehrer as a mixed growth of Wilms type, and Jung and Kroemer as a metaplastic chon- 
drosarcoma. The diffuse distribution of multiple islands of cartilage in edematous con- 
nective tissue of the ovary seems to favor a metaplastic origin. 

Sarcoma of Uterus. — Sarcoma of the uterus occurs in two rather distinct 
forms which it is desirable to separate because of their anatomical position, 
histogenesis, and clinical features. These are (i) mural sarcoma, and (2) 
sarcoma of the mucosa (Virchow). 

It was early recognized that certain apparently benign uterine tumors 
recurred after operation and these were described by English observers as 
recurrent fibroids (Hutchinson, Callender). In 1864 Virchow described a 
group of uterine sarcomas and these tumors have since been fully discussed 
by Gusserow, 1870, Tertillon 1890, Gessner 1899, Piquand 1905, and lately by 
Meyer, 1908. 

I. Mural Sarcoma. — This tumor occurs chiefly in the body, less often in 
the cervix (12 per cent. Gessner). It may appear as a primary sarcoma, or 
develop secondarily in myoma. In either case the growths present many of 
the features of uterine myoma. Thus they are often encapsulated, more 
frequently diffuse, and their position is subserous, interstitial, or submucous. 
The cervical and submucous tumors are usually polypoid and the interstitial 
growths may early break into the mucosa and thus become indistinguishable 
from originally submucous tumors. A rare type encircles the tubal orifice 
(Meyer). All varieties may be single or multiple. Multiple sarcomatous 
fibromyomas have been described by v. Kahlden, Gessner, and Busse. The 
single tumors may reach a very large size, and Terrillon has recorded a 
weight of 20 kilos. The submucous tumors and especially the cervical 
growths may be extensively polypoid or even papillary. 

On section the uterine sarcomas are soft and opaque as compared with 
myoma and are subject to liquefaction necrosis, and hemorrhage. Fatty 
degeneration, thrombosis of vessels and infarction, with extensive softening 
may produce large cysts in which suppuration may occur. While the mucosa 
is usually intact, many of the tumors which project into the cavity become 
ulcerated and extensively excavated. Overdevelopment of blood-vessels 
or lymph-vessels is observed and Meyer describes a markedly lobulated 
tumor in which the lobules were separated by large lymph-sinuses. The 
more circumscribed tumors have the general outhne of fibromyomas but 
diffuse forms infiltrate the wall, the mucosa, and the neighboring structures. 


Metastases appear late with encapsulated interstitial tumors but when 
the parametrium is invaded secondary growths often appear in the regional 
areas and in the lungs as well as in many other tissues and organs (Lubman). 
Yet Gessner observed regional metastases with circumscribed tumors. The 
uterine veins and lymphatics have been found thrombosed by tumor-masses 
without metastases. The ovaries may be extensively invaded through the 
vessels or by implantation. 

The gross anatomical features suggest the following somewhat serviceable 

1. Circumscribed primary sarcoma chiefly subserous or interstitial. 

2. Diffuse sarcomas chiefly submucous or infiltrating the wall and para- 
metrium, with metastases, and subject to hemorrhage and necrosis. 

3. Polypoid sarcomas of body or cervix. 

Fig. 72 — Polypoid sarcoma of cervix uteri, in a subject of 52 years. {After, McCann.) 

4. Extra-uterine single or multiple myosarcoma. 

5. Secondary sarcoma in myoma. 

Structure. — The structure of uterine sarcoma is extremely varied. Many 
authors have described them as myosarcoma, spindle-cell, round-cell, mixed- 
cell, and giant-cell growths and other subvarieties. Yet there is strong reason 
to believe that the entire group with rare exceptions, is of myogenic origin. 
It is well-known that the malignancy varies with the cell type, reaching its 
acme in round-cell and giant-cell structures, but it has repeatedly been shown 
that in the same tumor the structure of different portions may exhibit several 
cell types and that different tumors illustrate all transition phases of smooth 
muscle-cells to round-cells and giant-cells. Even in well-formed fibromyomas 
areas of round-cell sarcoma may occur (Meyer, Moraller, v. Franque). 
In a case of large polypoid submucous sarcoma, I found areas of adult myoma, 
very cellular lobules resembling the recurrent myoma, indifferent spindle- 
cell sarcoma, numerous areas of round-cells (very short spindles), and in 
edematous portions alveolar sarcoma with many large polyhedral cells. 
Moraller observed a notable perivascular grouping of round-ceUs. Meyer 
concludes that the round-cell sarcoma of cervix is often regarded as endothe- 



lioma. In fact it is evident that ia this field too much importance has been 

attached to variations in cell forms and too little to histogenesis and gross 


From the prognostic standpoint, how- 
ever, considerable significance attaches to 
the cell type and somewhat characteristic 
clinical varieties of uterine sarcoma are often 
associated with the predominance of cells 
showing different grades of anaplasia. 

{a) Recurrent myoma is a term that may 
well designate a tumor occurring chiefly in 
the cervix, which recurs^after a difficult ex- 
tirpation and which contains rather long, 
loose spindle-cells and scanty imperfectly 
formed long giant-cells. In a case of this 
type the structure remained uniform after 
three recurrences, while the pelvis and vagi- 
nal wall were slowly infiltrated. A charac- 
teristic recurrent myoma of the broad liga- 
ment occurs in the form of multiple nodules 
composed of weU-formed muscle-ceUs with 
marked vasoformative tendencies. The cells 
are short spindles with long nuclei and strongly 
acidophile cytoplasm. The arrangement of 
the cells is quite orderly, the vessels weU 
formed, and giant-cells, hemorrhage, degen- 
eration, necrosis, and metastases are wanting. 
{b) Large spindle-cells with granular cyto- 
plasm, large vesicular nuclei, and occasional 
giant-cells compose the bulk of many tumors 
which invade parametrium, broad ligament 
and mucosa, with ulceration, necrosis, and 
cachexia. In some cases the giant-ceUs are 
very numerous and of enormous dimensions. 
{c) Predominance of round-cells usually marks rapidly growing tumors 

which are locally destructive, produce cachexia from necrosis and hemorrhage. 

Fig. 73 •■ 

—Recurring myoma of 
cervix uteri. 

Fig. 74. — Section of a recurrent myosarcoma of vagina and cervix uteri. 

and early yield local and distant metastases (Meyer). Yet many metastases 
contain chiefly spindle-cells. Heinrich found both spindle-cells and round- 



cells in tubal metastases of spindle-cell tumors, and both original and metas- 
tatic tumors contained all transitions from muscle-cells to spindle-cells, 
giant-cells, and round-cells in the cases of Busse and Masty. 

Histogenesis and Etiology. — While many efforts have been made to incul- 
pate the connective tissue of the uterus in the genesis of sarcoma there is 
no sufScient evidence to show that these tumors include any of fibroblastic 

On the other hand the wide variation in form of neoplastic smooth muscle- 
cells demonstrated in these and other myosarcomas leaves little room for 
doubt that the entire group is essentially of myogenic origin. Yet there 
remains the possibility that the group of cells 
giving rise to the tumor contains besides 
muscle-cells a portion of the specific stroma- 
cells of the uterine mucosa. Such isolated 
complex groups of cells have been shown to 
give rise to adenomyoma, but the adenosar- 
comas which doubtless originate from such 
cell groups have quite a different structure 
from myosarcoma and contain gland alveoli 
(v. Winckel, Amann). 

In certain sarcomas areas of small round- 
cells traverse the main spindle-cell tumor in 
broad but sharply circumscribed bands and 
some have derived these bands from the stroma 
of the mucosa. Yet the adult stroma is 
probably not connected with the origin of 
any of these tumors and similar round-cells 
are observed in encapsulated intramural 

Regarding the nature of the embryogenic 
disturbance giving origin to sarcoma, one 
must refer to the same data which serve for 
uterine myoma. (For the sarcomatous trans- 
formation of myoma see myoma uteri.) 

2. Sarcoma of Uterine Mticosa. — Two 
main anatomical forms of this condition are 
commonly recognized, viz., (i) diffuse, and 
(2) polypoid. The latter represents a natural 
tendency of growth of the former, and there 
is no essential distinction between them. 

Arising from fundus or cervix the diffuse 
forms are usually widespread, Piquand find- 
ing the entire mucosa involved in 33 of 54 reported cases, while a portion 
or even the whole of the tubal mucosa may be affected (Griffith, Simp- 
son). In advanced cases polypoid outgrowths frequently -form. From a 
broad base bulky tumors may develop over which the uterus is stretched 
with thinning or hypertrophy of the wall (Beckmann, Terillon). Occlusion 
of the uterine canal may lead to retention of serous, bloody, or purulent 
fluids. In one case Terillon withdrew 1 5 hters of bloody fluid. A thin uterine 
wall may suffer inversion (Williams). The original mucosa may remain 
intact but is gradually altered by the growth of the tumor, by interstitial 
endometritis, or by erosion and replacement by granulation tissue. . 

The tumor is at first sharply separated from the musculans but soon 
invasion of the wall occurs and extension foUows to the parametrium, 

Fig. 75. — Myosarcoma o£ 
uterus. A rare tj^pe. The small 
cell groups surround minute vessels. 


peritoneum and vagina. Metastases are observed in the regional nodes, 
lungs, and elsewhere (Gessner). 

Secondary changes are common in both forms of sarcoma of the endo- 
metrium. The superficial portions early become eroded, infected, and ne- 
crotic. Hemorrhages, edema, myxomatous softening, wide areas of caseation, 
and fibrinous exudate combine to give a varied appearance. 

The polypoid form is most pronounced in the cervix where it is not readily 
distinguished from polypoid mural sarcoma and certain malignant teratoid 
growths. The teratoid tumors include the malignant myxosarcoma of 
Spiegelberg which is the most important cervical sarcoma and is to be sharply 
distinguished from the present group. Many reported polypoid sarcomas 
of the cervix refer to cellular polyps chiefly of inflammatory origin and it 
appears doubtful if the cervical mucosa is the seat of any true sarcomas apart 
from the two above types. 

The structure of endometrial sarcoma is quite varied. Ruge describes 
four varieties: i,i) Large round- and spindle-cells; [2) decidua-like cells; 
(3) giant-cells with small round-cells and spindle-cells, and (4) small round- 
cells. It is probable that all these forms represent variations of one original 
cell type. Thus Meyer found round-cells in the superficial layers, spindle- 
cells in the deeper portions of a pol}rpoid tumor. He interprets the decidual 
cells as transitory effects of the menstrual influence, and the giant-cells as 
regressive forms. In my own cases the chief features have been the diffuse 
arrangement of large and small rounded cells with hyperchromatic nuclei. 
The long spindles, parallel intertwining columns, and vasoformative tenden- 
cies of myosarcoma are wanting. The liability to necrosis is much more 

The diffuse distribution of many endometrial sarcomas is strong evidence 
that the tumor arises from the stroma-cells. Some have traced the earliest 
stages to the periacinar stroma. Wolgren describes as "fibromatosis uteri" 
a tumor-like thickening of the endometrium due to overgrowth of stroma- 
ceUs most marked about cystic glands, and Fellander places in an intermediate 
group between inflammatory and neoplastic processes his case of benign 
"elephantiasis fibrosarcoma tosa " in which many giant-cells were mingled 
with short spindles. The occurrence of such intermediate forms of hyperplasia 
renders it probable that highly malignant tumors also arise from the stroma- 
cells. That some of the supposed sarcomas of the mucosa are advanced 
stages of mural myosarcoma is nevertheless probable. Ruge states that 
sarcoma follows periacinar interstitial endometritis and that the tumor 
process first affects groups of glands about which the stroma-cells increase 
in size and number. 

For the endometrium a special factor in the genesis of sarcoma which 
should not be overlooked is the influence of gestation. Yet the varieties of 
sarcoma are well represented in children and virgins. Highly malignant 
lymphosarcomas are said to have existed in rare cases, but their exact 
nature is uncertain (Wihschamin, Wagner, Gow). Melanosarcomas result 
from imbibition of blood. Alveolar and angiosarcomas illustrate occasional 
secondary structural features. 

Carcinosarcoma of Uterus. — ^In a series of cases carcinoma and sarcoma 
have been found in the same uterus and the interpretation of these cases has 
given rise to much discussion (Gessner, Meyer, Herxheimer, H. Albrecht). 

I. Simultaneous occurrence of two separate tumors has been observed in 
ten cases and constitutes the most frequent and least notable form of the 
iCombination. Usually a polypoid sarcoma is associated with diffuse adeno- 
.carciuoma of the body (Niebergall). , 


2. Two tumors arise separately but later one invades the other (Nebesky, 

3. At the point where a submucous or mural sarcoma meets the glandular 
layer carcinoma develops secondarily. This event seems to be not infrequent 
and is illustrated in Albrecht's case where carcinoma developed in the mucosa 
of a uterus inverted by a sarcomatous polyp. It may indicate a stimulating 
influence of sarcoma-cells upon epithelium or, more probably, the ordinary 
effects of the hyperemia and irritation to which the mucosa becomes exposed. 
I have seen it in connection with a benign submucous fibromyoma, and near 
a calcific myoma. 

4. The glands included in sarcomatous polyps or in inflammatory polyps 
or in adenomyoma may become carcinomatous. In Gebhard's case the 
glands surrounded by round-cell sarcoma of the mucosa appeared to become 
carcinomatous. Amann's case has usually been interpreted as an adeno- 
sarcoma in which the glands became carcinomatous and appeared alone in 
the recurrences. Ballin assumed a simultaneous process in both stroma and 

5. The sarcomatous transformation of the stroma of a cervical carcinoma 
was assumed to have occurred in a case of Lindemann's, but the gross appear- 
ance of his specimen strongly suggests an ordinary cervical carcinoma with 
vesical fistula and with marked inflammatory cell changes. 

6. Spurious cases are undoubtedly recorded in which atypical proliferation 
of glands was interpreted as carcinoma (Klein), or endothelial proliferation 
was so mistaken (Riederer), or in which carcinoma, becoming diffuse and its 
cells assuming a spindle form, bore a certain resemblance to sarcoma. 

There is little doubt that many so-called carcinosarcomas result from 
the atypical growth of epithelial cells. Gessner refers all such tumors to 
this origin, pointing out that their metastases are usually carcinomatous. 
The exudation and inflammation accompanying tumors of the endometrium 
greatly favor atypical cell-growth the scope of which is, I believe, under- 
estimated. I have studied several tumors which appeared to exhibit sarco- 
matous and carcinomatous structure but have regarded them as unfavorable 
material for accurate observations on histogenesis. R. Meyer holds that 
sarcomatous transformation of the stroma of a carcinoma has not yet been 


The malignant tumors of bone constitute one of the most important 
departments of oncology. Their very frequent occurrence, local destructive 
tendencies resulting in loss of hmbs, rapid extension with loss of Hfe, and 
the difficulties in establishing the -prognosis render them a complex surgical 
problem. Arising from cells whose natural function is to produce the most 
stable of tissues, these tumors rank with the most rapidly growing of neo- 
plasms. Although often most clearly connected with a traumatic origin 
the mode of action of the trauma remains highly obscure, and while the 
neoplastic nature is usually most obvious, the general etiology of some of the 
' growths reveals complex relations with normal growth of bone and physio- 
logical regeneration and with nutritional, inflammatory, and infectious 

Finally the histological structure of this group of tumors while often 
highly characteristic is subject to wide variations so that many problems 
of histogenesis remain unsolved. 

Historical. — Although Astley Cooper recognized a distinction between 
periosteal and medullary tumors of bone, the older clinicians separated 
imperfectly between primary sarcoma, secondary carcinoma, and chronic 
inflammations of bone. Boyer first employed the term osteosarcoma and 
apphed it to certain malignant bone- tumors. The microscopical distinc- 
tions between sarcoma and carcinoma of bone were established by Lebert 
in 1845. Virchow very fully described the morphology of these tumors 
but did not separate them sharply from cancer. In 1849 Robin described 
two types of marrow-cells which he believed gave origin to bone-tumors. 
One of these, the myeloplacque, was described as a, large giant-cell with 6 to 10 
nuclei each with nucleoli and found chiefly in young mammals in the spongy 
tisssue and between the marrow and the inner wall of the bone. These cells 
appeared abundantly in some tumors. The other type of cells, the medullo- 
celles, were large round-cells with single nuclei and often made up the major- 
ity of marrow-cells. They also were found in tumors. It is clear that 
Robin did not fully distinguish bone-forming from blood-forming cells in 
the marrow. KoUiker identified the myeloplacques with osteoclasts and 
the medullocelles with osteoblasts. 

In 1853 Paget also recogjnized a variety of tumors containing large giant- 
cells which he identified with the giant-cells of marrow, and Gray described 
these cells as forming a well-defined layer where the marrow comes in contact 
with the bone in subjects under 20 years of age. In i860 Nelaton contributed 
an important monograph in which he emphasized the importance of 
Robin's cell groups and sharply distinguished between the malignant 
tumeur d, medullocelles and the benign tumeur d, myeloplacques. In 1879 
Gross reviewed the history and morphology of giant-cell sarcoma, empha- 
sizing against Billroth's opinion, the benign prognostic importance of the 
giant-cell structure and showing in four fatal cases that this tumor might 
become malignant. 

Gross' description of the origin, structure, clinical characters and treat- 
ment of bone-sarcoma stands to-day as the classic contribution on this subject. 



Robin's derivation of the giant-cells' from bone-forming cells of the marrow 
did not receive complete endorsement. Rindfleisch, Virchow, Wyss, 
Ziegler, and Hulke derived the giant-cells from hypertrophied bone-cells 
liberated during the absorption of bone (Gross). Wegner in Germany and 
Malassez and Monod in France and many later writers have claimed that 
these cells are vasoformative in nature and of endothelial origin. Many 
observers have shown that they are present in many forms of inflammation 
of bone (Rustizky, Fehr, Barrie). 

In recent years, studies in this field have dealt with the origin of the 
giant-cells, the relation of giant-cell sarcoma to inflammation, and the general 
etiology, prognosis, and treatment of these tumors. 

Classification. — Tumors occurring primarily in bone include those arising 
from bone-cells or osteogenic tissue and those arising from bone-marrow. 
Bone-cells give origin to tumors which retain more or less of the function 
of bone production and are properly called osteogenic, while the marrow- 
cells give origin to a very different class of tumors which are called myelomas. 
Malignant tumors of bone-cells, although differing markedly in many features, 
are essentially one and the same disease which may be designated as 
osteogenic sarcoma. The anatomical varieties of osteogenic sarcoma result 
from exaggeration of bone production as in sclerosing osteogenic sarcoma, 
or overgrowth of blood-vessels as in the telangiectatic type, or from the 
predominance of fibrocellular tissue as in the so-called periosteal sarcoma. 
Many osteogenic sarcomas combine all three of these features. The location 
of the growth is of minor importance, and yet periosteal tumors are usually 
fibrocellular, telangiectatic tumors either originate in the medullary por- 
tion of the shaft or soon involve the marrow cavity, and sclerosing sarcoma 
fills the marrow cavity with dense bone. The attempt to classify these 
tumors as periosteal or medullary is therefore unsatisfactory, since most of 
them involve marrow cavity, shaft, and periosteum. 

The benign giant-cell sarcoma of epulis type is not osteogenic but is 
essentially connected with the absorption of bone and usually of cancellous 
bone. Its position is usually medullary. Its structure may arise as a 
secondary process in osteogenic sarcoma. 

A peculiar and somewhat rare tumor of bone presents the structure of 
angio-endothelioma. It is not osteogenic, although it may arise from the 
blood-vessels of bone, or, more probably from the vessels of the marrow. 
It is described as an endothelioma. 

The myelomas, although destroying bone, have no relation to osteogenic 

The entire group of primary sarcomatous tumors of bone and bone-marrow 
may best be classified as follows : 

Primary sarcomatous tumors occurring in bone. 

1. Osteogenic sarcoma. 

a. Fibrocellular, chiefly periosteal. 

b. Telangiectatic involving marrow cavity, shaft, and periosteum. 

c. Sclerosing, affecting marrow cavity, shaft, and periosteum. 

2. Benign giant-cell sarcoma of epulis type. 

3. Myeloma, arising from bone-marrow cells. 

4. Endothelioma, or angio-endothehoma, arising from blood-vessels 
of bone or marrow. (Described under endothehoma.) 

Osteogenic Sarcoma. — {a) Periosteal Sarcoma. — The ends of the long 
bones, femur, tibia, and humerus are the chief seats, but the smafl bones of 
hands, feet and skull and many other portions of the skeleton are involved. 
The main gross feature of the periosteal sarcoma is the production of a tumor 



arising in the inner layers of the periosteum, extending along and inclosing 
the shaft, with a sharply marked border, causing a relatively bulky rounded 
fusiform or lobulated tumor lying outside the shaft which long remains intact. 
This position favors a bulky growth so that periosteal sarcomas commonly 
reach a large size, some of them attaining very large dimensions after a 
long period. The integrity of the shaft is usual but not constant, for the 
more cellular growths early penetrate and destroy the compact bone and 

Fig. 76. — Periosteal sarcoma completely traversing shaft as a telangiectatic structure. 

lead to spontaneous fractures. Even' thus the extramedullary tumor is usually 
more bulky than with corresponding central telangiectatic sarcoma. 

According to the consistence periosteal sarcomas are (i) soft, cellular, 
and highly malignant, or (2) firm and iibrous, or (3) irregularly ossified. En- 
capsulation by the periosteum is maintained for periods varying with the 
character of the growth but the more malignant tumors invade not only the 
shaft but the surrounding muscles, fascije, vessels, and eventually the skin. 



Tendon insertions offer considerable resistance to progress of the growth. 
The general appearance of these tumors supports the view that the peri- 
osteum is not passively displaced by an underlying growth, but, as MaruUaz 
finds, that the periosteum contributes to the growth and urTdergoes sarco- 
matous transformation. A thin, bony capsule may be laid down by the 

Fig. 77. — Periosteal sarcoma, combined with difEuse central sclerosing sarcoma. 

Incision reveals a solid tumor mass in which the shaft is fused with ossi- 
fying tumors or eroded by cellular growths. An expansion of the shaft 
as noted with benign central giant-cell sarcoma is missing. Fatty and mucoid 
degenerations, hemorrhage, and necrosis may greatly alter the appearance. 
Maceration by removing the soft parts may leave bony trabecule radiating 
out from the shaft, or a globular mass resembling an exostosis, or isolated 
bony areas may be removed entirely. Many of the large tumors, especially 
of the spindle-cell variety, contain cysts filled with serous fluid, or semifluid 



detritus, or blood. Although sometimes quite vascular periosteal sarcomas 
do not pulsate. 

All the histological varieties of bone-sarcoma originate in the periosteum, 
but the spindle-cell type is most frequent. Giant-cells are absent in many- 
cases, and always less numerous than in benign central sarcomas, being 
found usually in areas of dissolving bone. The tendency to calcification and 
ossification is very prominent. In One group, in which the cells are usually 

large and polyhedral, the matrix is 
chiefly cartilaginous. Some of the 
most malignant bone-tumors are of 
this type, and at times the large 
polyhedral cells grow diffusely while 
the cartilaginous matrix is missing 
or reduced to hyaline traces. 

{b) Telangiectatic Bone-sarcoma. 
— A considerable proportion of bone- 
sarcomas present an extensive de- 
velopment of blood-vessels sup- 
ported by more or less cellular 
osteogenic tumor-tissue, they early 
destroy the shaft, penetrate the 
marrow cavity for a short distance, 
and rupturing the periosteum extend 
widely through fascise and muscle. 
The exact origin of these tumors is 
not satisfactorily located but they 
seem to arise from the blood-vessels 
and bony tissue of the shafts of long 
bones. They are sometimes called 
central osteogenic sarcoma, but the 
majority of central tumors are not 
of this type. They are often con- 
fused with benign giant-cell sar- 
coma, which is a central medullary 
tumor, but is not osteogenic nor 
malignant. The structure of this 
telangiectatic sarcoma is occasion- 
ally seen over an entire tumor, but 
more often it is associated with 
other structural types, so that it is 
best interpreted as an exaggeration 
of the blood-vessel element in osteo- 
genic bone-sarcoma. 

The absorption of bone occurs 
rapidly and the line of absorption while ragged is rather sharp, suggesting 
that the growth has originated in the marrow, but the same type of absorp- 
tion occurs in vascular portions of tumors which are chiefly periosteal. 
Spontaneous fracture quickly results in many cases and may be the first 
symptom of the disease to be detected. 

Telangiectatic sarcomas of bone are osteogenic but the production of 
bone and even of tumor-tissue may be slight. The more solid portions of 
such tumors usually present irregularly ossified or only calcified strands, 
plates, or spicules of imperfectly formed bone. In some cases the periosteum 

Fig. 78. — Telangiectatic, periosteal, 
spindle and giant-cell sarcoma of femur. 
Perforation of shaft. 



is displaced outward and succeeds in laying down a thin shell of bone which 
crepitates on pressure. 

In the malignant bone aneurism, the vessels are widely dilated into 
irregular sinuses and the tumor-tissue is correspondingly reduced, or limited 
to an irregular layer disposed along the periphery. Formation of bony 
or osteoid tissue is also reduced and may be missing. The tumors may pul- 
sate and yield a bruit. 

The head of the humerus, upper end of tibia, and lower end of femur are 
the chief seats of malignant bone aneurisms. As a rule the tumor proves 
highly malignant and recurs locally or in the lungs after amputation. 

The relation of the malignant bone aneurism to cavernous angioma, 
benign giant-cell sarcoma, and simple blood-cysts of bone, has been much 

Fig. 79. — Malignant central osteogenic sarcoma of lower end of femur. 

discussed. The opinion expressed by Nekton, Volkmann, Lucke, and 
Gross, that the great majority of bone aneurisms are forms of sarcoma has 
been supported by recent observers (Gaylord, Nakayama, Bloodgood). 

It is not always easy but yet highly important to distinguish between 
malignant bone aneurism and very vascular forms of benign giant-cell 
sarcoma. The malignant disease is almost never found until it has destroyed 
much of the shaft of the bone, whereas the benign process long spares the 
shaft and expands so slowly that the periosteum is able to lay down a firm 
bone capsule about the expanding tumor. The great majority of simple bone- 
cysts are not connected with tumors. 


(c) Sclerosing Osteogenic Sarcoma. — The production of much dense solid 
bone occurs in portions of many bone-sarcomas but occasionally it forms the 
bulk of the tumor or is exclusively present, and since it is associated with a 
prolonged clinical course, there are both anatomical and chnical grounds 
for recognizing it as a specific form of osteogenic sarcoma. These tumors 
were once the subject of much debate (cf. Gross), being regarded by many 
authors as bone cancers, until Volkmann and Virchow established their 
true nature and termed them sclerosing osteoid sarcoma. 

In typical cases a large portion of the end of a long bone is converted into 
a solid mass of dense almost ivory-like bone, which obliterates marrow cavity 
and shaft and produces a fusiform swelling of considerable dimensions. 
Blood-vessels are very scanty except about the periphery or over a small 
area of the tumor, yet metastases form in a high proportion of cases, so that 
this slowly advancing tumor has long been known to be fully as malignant 
as the others. They show comparatively little tendency to invade surround- 
ing soft parts. 

Areas of sclerosing sarcoma occur frequently with periosteal growths 
in which they show all grades of bone formation up to the very dense ivory- 
like type. 

Capsular and Parosteal Sarcoma. — Osteogenic tumors not infrequently 
involve the capsules of joints or appear in the soft tissues about joints or 
along the shafts of bones. This group of cases is important not only for 
their clinical interest but also for their bearing on the origin of bony tumors 
and of certain neoplasms not always recognized as related to bone. The 
capsule of the knee-joint is the chief seat of capsular sarcoma and here the 
tumors illustrate most of the clinical and histological features of bone-sarcoma 
(Rydygier, Lit.). They are distinguished chiefly by the failure to involve 
the joint-cavity or the bone, and their malignancy is distinctly less than 
that of corresponding tumors of periosteum or medulla. Yet malignant 
cases occur which are rapidly fatal after resection or amputation. Some 
produce a fusiform swelling of the joint and being accompanied by fever the 
case resembles tuberculosis (Garre). Or a localized tumor appears at one 
point in the capsule (Marsh). 

Pedunculated giant- and spindle-cell sarcomas have been described by 
Weir and by Dowd. The structural variations cover most of the scope of 
bone-sarcoma with corresponding prognosis. 

Capsular sarcomas of the ankle exhibit much the same characters as in 
the knee (Moser, Lit.). The tendon-sheaths are a common seat of sarcoma 
which usually takes the form of a giant-cell tumor of slight malignancy. 
The deep fasciae may be the seat of sarcomas of various osteogenic varieties 
which have no definite connection with the periosteum and the true nature 
of which may therefore be overlooked. They include giant-, spindle- and 
round-cell types, and may belong in the group of chondrosarcoma or its 
derivative myxochondrosarcoma. 

Along the cranial sutures, base of skull, pharynx, nares, vertebrae and 
pelvis, occur many tumors of pronounced osteogenic character which belong 
in this group, and not a few in which the derivation from bone is only vaguely 
indicated by the structure. 

The course of osteogenic sarcoma varies greatly and accords in general 
with the histological structure. As a rule they are malignant tumors appear- 
ing before the 30th year, recurring after incision or amputation and producing 
pain, fever, metastases, and cachexia. Coley reports four bone-sarcomas 
occurring in subjects aged 17 months to 9 years, and Goebel has collected 
cases of congenital bone-tumors. The spindle-cell tumors grow rather 


slowly but nevertheless tend to recur and cause death by local growth and 
internal metastasis. Jackson recorded an unusual case in which a small 
growth on the under side of the knee increased very slowly for 32 years when 
after trauma it grew rapidly and produced a large medullary tumor. The 
highly osteoid tumors may also pursue a very chronic course, as with cases 
of Paget and Holmes which proved fatal with metastases after 21 and 25 
years. j 

Metastases are frequent with osteogenic periosteal sarcoma, in the ad- 
vanced stages of all types and very early with soft and cellular varieties. 
A period of months or years may intervene between the appearance of the 
primary and the secondary tumors or there may be almost simultaneous 
involvement of several bones (LeDentu, Zahn, Poncet). That these tumors 
are not myelomas is attested by Poncet's case. The lungs are chiefly in- 
volved by emboH through the blood-vessels and other widely distant organs 
may be reached by the same route. Neighboring or distant lymph-nodes are 
also occasionally invaded through lymphatics but swollen lymph-nodes 
usually prove to be purely inflammatory. The occurrence of pulmonary 
tumors secondary to periosteal sarcomas was noted by Virchow but the 
frequency and extent of such metastatic tumors has lately been emphasized 
byLeCount (Lit.). 

These tumors appear as multiple nodules, or as superficial pleural growths 
or as bulky pleural or pulmonary tumors. Extensive diffuse pleural growths 
are reported by Virchow, Pitts, LeCount and others, and pedunculated 
pleural tumors by Virchow and Cocks and Wilks. Massive intrapulmonary 
growths have been observed by Birch-Hirschfeld, Kuster, AUin, Meakin and 
others. In structure the tumors have been soft and cellular or cartilaginous 
and bony, or encased in a bony shell. In a remarkable case described by 
Bosch, the pulmonary metastases of an osteoid sarcoma of the femur ex- 
hibited lamellar structure as in the long bones, and enclosed groups of cells 
resembling lymphoid marrow. Pleural effusion may be absent and pulmon- 
ary symptoms may be slight, but bloody effusions, dyspnea and hemoptysis 
frequently occur and suggest tuberculosis. 

In a few cases extensive intravascular growths have been observed. 
Hektoen found bony masses from a tibial sarcoma, which nearly filled the 
right ventricle and produced secondary growths in the lungs. Feistmantel 
found cylindrical bone masses in one lung and large nodules in the opposite 
lung which probably resulted from metastases in the pulmonary artery from 
a femoral sarcoma. Pulmonary metastases appear to have occurred while 
the primary tumors, of foot and wrist, were quite small in the cases of Pitts 
and of West. Usually they appear a few months after operation on the 
primary growth, while in j^enckel's case death from a pulmonary growth 
occurred 15 years after amputation of the femur. 

That metastases are often produced by the trauma of operation is in- 
dicated by McAuliffe's observation of tumor-cell emboli in many organs 
after death from resection of a sacral sarcoma. 

Benign Giant-cell Sarcoma (Epulis Type) . — This peculiar condition was 
first fully described as a benign process in bone by Nelaton, who emphasized 
the reddish jelly-like appearance of the tumor-tissue, noted its capacity to 
absorb bone and widen the marrow cavity, and fully established its benign 
course and the wisdom of conservative treatment. All of these features 
have been full recognized by most later observers but not by all, so that many 
Hmbs have been needlessly sacrificed because of the failure to distinguish be- 
tween this benign process and the malignant central telangiectatic sarcoma. 

The disease appears at the ends of the long bones, but it very frequently 


arises in the maxillae after the extraction of teeth so that it is often called 
the epulis type of sarcoma. It has frequently appeared as a multiple tumor, 
and one of the most remarkable of these cases is that of Martland in which 
tumors appeared in widely distant parts of the body over a long period 
without showing any evidence of local malignancy. 

Very similar tumors occur in the capsules of the joint, and along the 
tendon sheaths and bursas. 

In the long bones the tumors produce reddish jelly-like tumor masses 
resembling granulation tissue and replacing the cancellous portions of the 
bone. They may also slowly absorb the shaft while the periosteum lays 
down an advancing shell of new bone. Eventually this shell may become 
thin and allow passage of the tumor-tissue but there is seldom any tendency 
toward invasion of soft parts. Likewise the cartilaginous surfaces of the 
joint may be reached and absorbed and the joint surfaces may collapse, from 
simple absorption but without infiltration (Gross, Stimson). 

There is considerable variation in the rate of growth and bone absorption 
by benign giant-cell sarcoma. As the spindle-cells of the stroma become 
more active and abundant, the giant-cells diminish, the tumor shows less 
resemblance to granulation tissue, but becomes firmer and like spindle-cell 
sarcoma. Such growths recur locally and may destroy bone and infiltrate 
the soft tissues, but I have never known them to yield metastases. They 
regularly fail to produce bone. 

The earliest stages of this disease have been traced to the walls of small 
cysts forming in osteitis fibrosa cystica. Here the process may be interpreted 
as a form of bone absorption, or as some prefer, rarefying osteitis or osteomye- 
Htis, with the formation of cysts. 

In some cases the process appears to subside, leaving cysts with smooth 
lining and serous contents. In most cases it is progressive and acquires the 
characters of a mild tumor process, and rarely it develops greater momentum, 
actively destroys bone and exhibits certain clinically malignant tendencies. 

Structure of Bone Sarcoma. — There are three main histological varie- 
ties of bone-sarcoma: (i) spindle-cell, (2) giant-cell, and (3) round-cell tumors 
and in each of these classes there are marked variations which are of 
importance in prognosis. Since all these tumors arise from bone-forming 
cells the varying morphology may be interpreted as the expression of different 
grades of anaplasia. To some extent also it probably represents the varying 
etiology, and the complex pictures resulting from bone absorption and imper- 
fect bone production. Inflammation, degeneration and necrosis also leave 
pronounced effects in many cases. 

I. Spindle-cell tumors are most frequent and this is the usual form of 
periosteal growths. The cells have a short spindle form, or they appear 
large and almost polyhedral, and in very malignant processes they are small 
and appear almost round although never assuming lymphoid characters. 
They present many variations in size and chromatin content of the nuclei 
which are of much importance in prognosis. Intercellular substance varies 
from a scant trace in soft mahgnant growths to the great abundance of 
osteoid or bony tissue in the osteoid sarcomas. 

The slowly growing sarcomas are poor in vessels but the malignant cellular 
forms may be very vascular. The walls of the larger vessels are often infil- 
trated with round-cells, occasionally they appear sarcomatous, and rarely 
they have been found distended with tumor-tissue (Bristowe). 

In vascular tumors a close relation of the tumor-cells to the blood-vessel 
is common and occasionally the structure of perithelioma is simulated. A 
pseudo-alveolar structure is sometimes observed. 



A distinguishing character of the spindle-cell periosteal tumor is the 
production of bone or cartilage. The bone appears in short trabeculae dis- 
tributed throughout the growth, or more often in the form of osteoid tissue 
which may compose the bulk of the tumor. In the osteoid tumors the cells 
are reduced in number and appear partly or completely inclosed in lamellae 
of hyaline partly ossified material. Areas of cartilage are often mingled 
with the osteoid tissue and many phases appear of the transformation of 
cartilage and osteoid tissue into bone. Most tumors, however, fail to carry 
the development of bone to a completion. The most advanced bone pro- 
duction is seen in the stalactite tumors which resist maceration, and in the 
diffuse ivory-like growths. When osteoid sarcoma invades the muscles the 
histological picture may resemble myositis ossificans. When bone is ab- 
sorbed by the progress of the sarcoma, the solid trabeculse first exhibit inter- 




Fig. 8o.- 

-Periosteal sarcoma. Vein lined with tumor cells. This structure explains the 
ready metastases o£ these tumors. 

fibrillary clefts from advancing decalcification, then canals are formed by 
advancing blood-vessels or groups of tumor-cells, or giant-cells form in 
Howship's lacunae, producing the lacunar resorption. Likewise when an 
osteoid sarcoma suffers resorption its structure changes and all the above 
processes may be observed. In all cases where bone, normal or neoplastic, 
is being absorbed the giant-cells tend to assume the foreign body type of 
benign sarcoma. 

2. Giant-cell Sarcoma. — When giant-cells predominate or form a promi- 
nent feature of bone-sarcoma the tumor is called a giant-cell sa,rcoma. This 
term has been applied chiefly to sarcomas of the epuhs type, central or 
periosteal, in which characteristic giant-cells are abundant or form the bulk 
■of the tumor, and it would be advisable to restrict the term to this relatively 




benign tumor. Similar cells are sometimes seen in portions of malignant 
tumors. These, cells are of large size, the cytoplasm is opaque and acido- 
phile, and the nuclei are numerous, separate, oval and usually grouped in 
the center of the cell. They often contain vacuoles, fatty detritus, homo- 
geneous material, red blood-cells, and occasionally spicules of bone. They 
have the general characters of foreign body giant-ceUs (Fig. 82). 

The bulk of these growths is composed of loosely packed spindle-cells, 
surrounding many vessels, capillaries, and sinuses. There is much variation 
in the vascularity of these tumors and in the number, size, and nuclear chro- 
matin of the spindle-cells, and there is 
a corresponding difference in their 
powers of growth. 

Periosteal sarcoma may contain 
giant-cells of the above type, espe- 
cially where the tumor comes in contact 
with absorbing bone, but both peri- 
osteal and central sarcomas may exhibit 
giant-cells of distinctly different type 
and significance. These cells may be 
large or small, the cytoplasm stains 
poorly, the nuclei are large and vesi- 
cular and composed of many lobes which 
may become detached. As a rule 
these cells belong to malignant tumors. 
While the two types of giant-cells are 
usually distinct, it must be admitted 
that transitional forms seem to occur, 
especially when mahgnant tumors un- 
dergo regressive changes, erode bone or 
become inflamed (Fig. 83). 

The existence of dififerent types of giant- 
cells in bone-sarcoma has been recognized by 
several observers. Poncet describes true and 
false myeloplacques. The former are as large 
as 3 to 4 ordinary tumor-cells and contain s to 6 
nuclei; the latter are as large as 4 to 5 ordinary 
cells and contain 12 to 20 nuclei, and corre- 
spond to the common giant-ceU of the epulis. 
Borst states that the giant-cells of medullary 
sarcoma are more delicate than those of peri- 
osteal origin, in which the cytoplasm is more 
compact, often free from granules, and more 
sharply defined, and the cell processes are longer than in medullary tumors. Mallory dis- 
tinguishes between true tumor giant-cells and the foreign body giant-cells of osteogenic 
and other sarcomas. The tumor giant-cells differ from other tumor-ceUs chiefly in size 
and number of nuclei which may multiply by mitosis. They produce specific fibrils. 
The foreign body giant-cells are of endothelial origin, and correspond in type to those of 
the epulis. 

The origin of the giant-cells of the epulis type has been the subject of much discussion. 
Virchow and Rindfleisch believed that they were hypertrophied bone-cells set free by the 
absorption of the bone matrix and that they are identical with the osteoblasts of Kolliker. 
Robin and Nelaton, in describing the giant-cell sarcoma of bone (tumeur a myeloplaxes), 
accepted the identity of the giant-cells with KoUiker's osteoblasts. This view is favored 
by the very frequent presence of giant-cells in many lesions of bone and has been widely 
accepted. The direct transformation of bone-cells into giant-cells of sarcoma has been 
traced by Wyss, Ziegler and others, but it is not clear that the giant-cells observed in 
the mahgnant tumors of those authors were identified with those of the epulis. Ziegler 
found that giant-cells are by no means constant in the absorption of bone and suspected 
that osteoclasts might have a variable origin. 

Fig. 8i.- 

-Malignant" chondrosarcoma of 



Wegner, Malassez, and many later observers have been led to conclude that the giant- 
cells of the epulis are derived from modified endothelium or angioblastic cells, and Malassez 
classiiied this tumor with the angioblastic sarcomas. Very similar results were reached 
by Ritter who traced the cells in Howship's lacunae to proliferating endothelium and he 
further concluded that these cells may take on the function of bone-cells. Borst recognizes 
the multiple origin of the giant-cells of sarcomas, but clings to the view that in the epulis 
they stand as evidence of the bony origin of the tumor matrix. Mallory holds that the 
giant-cells of epulis are transformed wandering endothelial leukocytes and do not belong 
to the tumor process. 

Thus, both in morphology and in origin it is necessary to recognize two 
types of giant-cells in bone-sarcoma. One occurs chiefly in the benign 
epulis, is chiefly of endothelial origin, and belongs among the foreign body 

Fig. 82.— Benign giant-cell sarcoma of head of tibia. Epulis type of medullary sarcoma. 
Giant-cells forming about blood masses and lipoid material. 

giant-cells, although participating in the tumor process. The other occurs 
chiefly in malignant periosteal tumors and is derived from tumor-cells. 

The giant-cells of bone-sarcoma are therefore contributed both by the 
bone-cells and by proliferating angioblastic cells. Apparently the growing 
endotheUum feels the influence of the mother tissue and may take on some 
of the form and function of bone-cells, giving the specific structure of the 
benign epuhs which is rarely observed except in connection with bone. 

On account of their chiefly angioblastic origin and benign course many 
have suggested that bone-tumors of the epuhs type should be excluded from 
the class of sarcomas or even of true tumors. Yet these cases show many 
features of neoplasms, not a few are moderately mahgnant, and many transi- 
tional forms up to highly mahgnant tumors are observed. 



3. Round-cell Sarcoma of Bone. — In many very malignant bone-tumors 
the cells are of small size and assume a short spindle or rounded form. Since 
these tumors are not of lymphoid origin and since traces of the spindle 
form are nearly always present it is of doubtful wisdom to employ for them 
term "round-cell." Moreover, there is a true round-cell tumor of the bones 
which is of lymphoid character, the multiple myeloma, and which should be 
sharply distinguished from osteogenic sarcoma. Yet in rare cases of perios- 
teal sarcoma groups of lymphoid cells may be found which seem to represent 
abortive efforts toward the production of bone-marrow. At present it is 
impossible to state to what extent cells of this character enter into the struc- 
ture of the common bone-sarcomas. 

Fig. 83. — Periosteal sarcoma. Malignant giant-cell structure. 

These tumors are the most anaplastic actively growing and malignant 
of bone neoplasms. They rapidly destroy the bone, periosteum and marrow 
cavity, infiltrate joints, and surrounding tissues, recur promptly after opera- 
tion and produce early metastases. 

In the common small cell tumor the cells are of short spindle form, or 
on cross-section round or polyhedral. Multinucleated cells are often pres- 
ent. Mucinous and fatty degeneration, hemorrhage, and necrosis are 
frequent in this variety. 

Prognosis of Bone-sarcoma. — The widest extremes are represented in 
the prognosis of different sarcomas of bone, and it becomes a matter of first 
importance to determine the prognosis in the light of all the data, clinical 
and pathological. 

The age of the patient, the duration of the disease, its rate of progress, 
the particular bone affected, the completeness of the operation, and above all 


the structure and essential malignancy of the tumor, must all be considered. 
Subjects under 20 years withstand the progress of malignant tumors poorly. 

The duration of the disease must be considered from several standpoints. 
A tumor which grows very slowly is for that reason probably less mahgnant, 
yet there are numerous records of long existing tumors suddenly assuming 
active growth after trauma (Jackson). Certain slowly growing but malig- 
nant osteoid tumors appear not formidable in themselves but early produce 
metastases (Le Count). The local damage to tissues occurring in the 
advanced stages of otherwise comparatively benign tumors is an impor- 
tant element in prognosis, chiefly because it complicates the operation, but 
also because it favors invasion of veins and metastasis. Kramer states 
that encapsulation is the most important factor in the prognosis of most 

The rapidity of growth is a very reliable chnical sign of malignancy and 
should always control the deductions drawn from histological structure. 
Rapidly growing tumors are nearly always malignant and eventually fatal. 
A highly anaplastic, small spindle-cell, traumatic sarcoma of patella sub- 
mitted to me by Dr. John Rogers proved fatal six weeks after the trauma, 
with very extensive local recurrence and metastases. Many would have 
described this growth as of round-cell structure. 

Yet the essential capacity for growth cannot always be estimated clin- 
ically, encapsulation restrains growth, and it is possible for a tumor to change 
its rate of growth. A central tumor which long remains confined by the 
periosteum usually has a limited growth capacity, for very malignant tumors 
early perforate the shaft. Yet in one of Bloodgood's fatal cases symptoms 
had existed for 25 years. 

It is an established rule that periosteal growths are much more malignant 
than central tumors but this rule has little practical value. Gross found the 
average duration of life with round-cell, spindle-cell and osteoid periosteal 
tumors with operation at 18, 20, and 92.7 months respectively. While 
amputation in the early stages of spindle-cell or osteoid sarcoma of 
the extremities may be followed by complete recovery, the prognosis of any 
variety of periosteal sarcoma located in or near the trunk is very grave. The 
great majority of cases which seem to recover after operation eventually 
suffer from local recurrence or internal metastasis. Gross concluded that 
practically all cases of round-cell or spindle-cell periosteal sarcoma eventually 
.die of 'the disease, while with osteoid sarcoma 65.62 per cent, of the cases 
prove fatal. More recent statistics show slightly better results and some 
dependence of the outcome upon the particular bone affected. 

Kocher collected 48 cases of sarcoma of long bones which were regarded 
as cured. Of these the radius was the seat of the growth in 4, ulna i, humerus 
10, femur 23, tibia 12. Coley collected 57 cases of apparent cures, 30 of 
the "myelogenous" type, 15 periosteal, 12 undetermined. Butlin collected 
18 cases of periosteal sarcoma of humerus which survived operation, but 
only one remained well after three years. Of 68 cases of periosteal sarcoma 
of femur only one doubtful case remained well three years. Yet McCosh 
had three recoveries of spindle-cell sarcoma of the femur among seven cases. 
Bloodgood reports 26 medullary giant-cell sarcomas all living; 18 inoperable 
malignant sarcomas; 34 high amputations without a single cure; and 6 
permanent cures (periosteal osteosarcoma, 3, myxochondrosarcoma, 2, 
fibrosarcoma, i). 

Rheinhardt reported 7 permanent cures (8 to 12 years) out of 54 cases, 
including 4 round-cell tumors, of tibia 2, and humerus, 2. Coley collected 
62 cases of sarcoma of clavicle of which 6 recovered but in many the result 


was not known. Nancrede could find no cases of recovery from sarcoma 
of the scapula. The interscapulothoracic amputation, while comparatively 
effective for benign tumors of the humerus, appears to have succeeded in 
curing only one advanced malignant tumor of the humerus (Berger's case) 
(Jeanbreau, Riche). 

From a general survey it appears that recovery has occasionally followed 
operation for almost every variety of sarcoma of the long bones, but in none 
of the statistical reports of Nasse, Reinhardt, Koeber, Buthn,_ Coley, has 
adequate attention been given to the dependence of prognosis upon the 
location, origin, and structural type of the tumor. 

Although Billroth asserted that the histological structure is of no aid 
in prognosis this opinion cannot apply to the more mahgnant tumors, most of 
which prove fatal and among which recovery is determined by the location 
and stage of growth of the tumor. Contrary to the above opinion histo- 
logical studies have estabhshed the benign nature of the central giant-cell 
sarcoma, the greater maUgnancy of other types of giant-cell tumors, and the 
very unfavorable prognosis of the small spindle-round-ceH growths. The 
chief source of difl&culty and confusion in this field has arisen from the as- 
sumption that the tumors fall into a few sharply defined categories, as giant-, 
spindle-, and round-cell, whereas there are different grades of potential 
malignancy in each of these groups. When this factor is considered I 
believe it will be found in bone-sarcoma as with other groups of tumors that 
there is a definite relation between prognosis and structure. 

1. The benign nature of central giant-ceU sarcoma of the epuhs type 
was early pointed out by Nelaton, later by Gross, Haberer, and others, 
and is now generally recognized. Bloodgood finds that practically all of 
these tumors respond to conservative operations, and that considerable 
invasion of the soft parts is no bar to conservative treatment. This treat- 
ment consists of curettage, or resection, followed by appUcation to the wound 
of pure carbohc acid, and alcohol, or chloride of zinc. Implantation of 
bone is employed when necessary. Among 26 cases curetted, Bloodgood 
records five recurrences all of which were cured- by a second curettage or 
by resection or by amputation. Kramer and Hinds report successful 
curettage when both condyles of the femur were involved. Most of these 
tumors respond well to radium. 

It was early shown that certain giant-cell sarcomas of bone are very 
malignant (Ziegler, Wyss, Gross), and this fact has stood against the general 
recognition of the benign nature of the epulis type. Yet the giant-cells 
of the mahgnant tumor are quite different in appearance from those of the 
epulis and present single multilobed hjrperchiomatic nuclei while the chief 
spindle-cells of the tumor also show the features of malignancy, mitoses 
and hyperchromatic nuclei. Gross found that the mahgnant giant-cell 
sarcomas exhibit calcific deposits and bony trabeculae, but these features 
are wanting in the sarcomas of the epulis type. 

2. The telangiectatic sarcoma and the bone aneurism usually contain 
many giant-cells but the prognosis of these tumors varies extremely, some 
proving highly mahgnant. Here I have several times observed that 
the character of the giant-cells is not a safe guide but that in the mahg- 
nant cases the chief or spindle-ceUs present unmistakable features of maUg- 
nancy. My experience leads me to conclude that telangiectatic bone- 
sarcoma tends to prove mahgnant when it occurs near the trunk, when 
round- or spindle-cells with hyperchromatic nuclei predominate over giant- 
cells, and when perforation and destruction of the shaft occur early. In the 


malignant cases reported by.Nakayama and Bloodgood the scanty tumor- 
tissue was composed chiefly of round- and spindle-cell tissue. 

3. A source of error arises when one portion of the tumor differs from 
another. When ulceration and necrosis occur the superficial portions of 
the _ tumor may lose their original structure and giant-cells assume the 
benign foreign-body type. I have made an erroneous prognosis of benign 
sarcoma from the fungating tissue about a sinus leading to a mahgnant, 
sclerosing, central and periosteal tumor. 

4. With spindle-cell growths considerable differences in structure and 
mahgnancy exist which are .constantly revealed by the minute characters 

1*^ .-(--^ ^tr cJ* • ." 




Fig. 84. — Malignant telangiectatic central sarcoma of head of humerus in a child. 
Malignant bone aneurism. The giant-cells are of the epulis type, but the malignancy is 
indicated by the spindle-cells with very hyperchromatic nuclei. 

of the cells. Comparatively adult types of spindle-cells with much fibrous 
or osteoid tissue belong to the slowly growing tumors which, however, 
recur locally and in advanced stages metastasize. Here the prognosis depends 
upon the location, extent, and complications of the tumor. At the other 
extreme stand the very malignant rapidly growing small spindle-cell tumors, 
often called round-cell sarcoma, for which a fatal prognosis must be given 
at all stages. Tumors of large polyhedral or rounded cells occur in bone 
which in general exhibit marked malignancy. The exact grade may be 
estimated by the minute characters of the cells, especially by the activity of 


cell division and by nuclear hyperchromatism. Fibrosarcoma and cellular 
chondromas are comparatively benign. Myxochondrosarcoma is also 
as a rule not very malignant, and if as is often the case such tumors are 
parosteal and encapsulated, the condition invites conservative treatment. 

5. It should be emphasized that the histological structure indicates only 
the potential malignancy of a tumor, or its rate and capacity of growth. 
This factor is only one of many which determine the actual course of the 
tumor, which forms the clinical conception of malignancy. In the field 
of bone-sarcoma when the chnical course does not accord with the structural 
characters the discrepancy should be attributed to lack of information 
regarding the gross and clinical features of the case. 

Etiology. — The etiology of bone-sarcoma is highly obscure. While 
trauma must be accepted as a very common exciting factor, little is known 

Fig. 85. — Central sarcoma of bone. (Epulis type.) This particular tumor is more cellu- 
lar than usual, and is not entirely benign. 

about the growth or structure of bone which can explain the mode of action of 
the trauma. A study of the healing process in fractures is yet very sug- 
gestive of the mode of origin of traumatic sarcoma. In the average specimen 
of callus the proliferative activity of fibroblasts, osteoblasts, and endothelium 
is quite remarkable and often presents a picture which is difficult to separate 
from sarcoma. Especially in certain foci the cells may be quite as abundant 
as in sarcoma, but their arrangement is more orderly and the nuclei are not 
hyperchromatic. The organization of considerable masses of blood provides 
conditions favorable for very free growth of atypical cells. Some organizing 
blood-clots after fracture, I have found quite difficult to separate from 
sarcoma. Formation of progressive traumatic angiornas which gradually 



extend beneath the periosteum, often inclosed by a shell of new bone, may 
apparently precede the late development of traumatic sarcoma. 

As a rule the trauma is moderately severe and ma}' result in fracture or 
splintering, or in local hemorrhage, and the tumor seems to represent- an 
exaggeration and perversion of the healing process. Klebs describes a very 
early medullary sarcoma in which the adventitial cells of the blood-vessels 
seemed to be chiefly concerned. There is some evidence that the presence of 
blood exerts a peculiar stimulating influence on the regeneration of bone. 
Both these factors are present in the healing of fractures and other traumatic 

All varieties of sarcoma have been attributed to trauma, but it is evident 
that the injury is only one of several essential factors, since many or 'most 
cases fail to give a history of trauma, many tumors arise in bones protected 
from injury, and occasionally bone-sarcomas are multiple. The traumatic 

Fig. 86. — Structure of malignant chondrosarcoma of 

; bone. 

origin applies chiefly to periosteal sarcoma. The fact that central sarcomas 
arise usually near the epiphyseal hne indicates that disorders in the growth 
of bone at this point are involved in their origin. Borst reports a remarkable 
case of periosteal ossifying sarcoma which contained islands of lymphoid 
cells and hematoblasts, and he suggests that the tumor arose in part from an 
aberrant portion of marrow-tissue. For the common epuhs of the alveolar 
border injury to the periosteum and bone from extraction of teeth is usually 

The origin of parosteal sarcoma must be referred to misplaced islands 
of bone-forming tissue. An imperfect differentiation of tissues neighboring 
to bone such as has been assumed to exist in myositis ossificans may also 
be included as an etiological factor. The observed facts do not seem to 



require the theory of a metaplastic development of any recognized type of 

Simple chronic inflammation of various types may be concerned with 
some of the capsular sarcomas. Syphilis and tuberculosis may be suggested 
as probable factors in certain of the round-cell and diffuse sarcomas but 
the grounds for such an assumption are quite indefinite. The Wassermann 
reaction 'has been present in some of. my cases of extensive bone-sarcoma. 

The idea that trauma or any other factor may lead to the development of 
sarcoma at the ends of long bones which are previously entirely normal is, 
I think, without satisfactory foundation. There is, however, abundant 
evidence that developmental disturbances, chiefly at the epiphyseal lines, 
form an essential factor. The routine examination of the ends of bones in 
infants shows not a few irregularities in growth resulting in the displacement 

Fig. 87. — Excessive callus developing in two weeks about a comminuted fracture of inner 
end of clavicle. Clinically resembling sarcoma. 

of islands of cartilage in marrow cavity or periosteum. These irregularities 
I find especially frequent after rickets. Many osteogenic sarcomas are 
sharply separated from the marrow cavity by a septum of bony or osteoid 
tissue, indicating that the tissue of origin was also separated. After rickets 
definite massesof imperf ettly ossified cartilage are often found in the marrow 
cavity. It is reasonable to assume that mild forms of bone dyscrasiae exist 
which are not recognized clinically but which are capable of causing abnor- 
malities in the formation of the ends of the bones with a tissue predisposition 
to tumor growth. 

The elements of tissue predisposition and constitutional tendency become 
most apparent in connection with the disease osteitis fibrosa. 

Osteitis Fibrosa.^This interesting malady, first fully described in 1891, 
has important bearing on the etiology and significance of giant-cell sarcoma of 
bone of the epulis type (Haberer, Bloodgood, Lit.). It occurs chiefly in chil- 
dren and has been termed osteodystrophia juvenilis (Monckberg, Tietze, 


Konig) . It also occurs in elderly subjects where its relation to osteomalacia is 
apparent. It pursues a long course but often ends fatally from cachexia or 
intercurrent disease. It is characterized by defective calcification of the 
bones which leads to thickening, weakening and deformity of one or many 
bones of the limbs, jaws, skull or trunk. The bone tissue is replaced by cel- 
lular fibrous tissue which may invade the marrow cavity replacing the red 
marrow. From this stage the disease takes one of two courses, the cellular 
tissue softening and producing multiple bone cysts lined by fibrous tissue and 
filled with clear fluid, or in the walls of the cysts or with cyst formation giant- 
cell sarcoma develops. There has been much controversy regarding the rela- 
tion of the different stages of this process, some holding that the cysts always 
follow the sarcomatous process, others finding that the sarcoma develops in 
the cysts. Both events seem to occur. A history of traurr a and the presence 
of hemorrhage and pigmentation in the' walls of cysts seem to show that the 
presence of blood coagulum influences the development both of cysts and of 
sarcoma. In some cases fibrous cysts are chiefly observed (Monckberg, 
Heinecker) in others cystic sarcomas (Haberer) or again solid sarcomas 
(Rehn). The long duration, nine years (Rehn) and relatively benign quality 
of the sarcomatous process are notable features. Arising in the course of a 
nutritional disturbance of bone and being associated with inflammatory 
changes Rehn concludes that the entire process is inflammatory. Cystic 
softening of the tumors, transformation into fibrous tissue, and spontaneous 
disappearance have several times been observed. Lubarsch points out that 
very similar processes giving rise to a diagnosis of sarcoma occur at the points 
of fractures in cases of osteitis fibrosa and he inclines to the opinion that the 
process is not a true sarcoma (Gaugele). 

A remarkable case reported by Martland of multiple giant-cell sarcomas 
involving many of the bones of the body probably belongs in this category 
and illustrates a universal tissue predisposition to the disease. Knaggs and 
Gruner describe an atypical case of fibrous osteitis affecting femur, tibia, and 
fibula, and ending in malignant sarcoma. 

Kolisko identifies osteitis fibrosa with Paget's osteitis deformans and finds 
that in both the transformation of bone-marrow into fibrous tissue may be 
followed by regressive changes leading to bone cysts or to progressive changes 
giving rise to giant-cell sarcoma of slight malignancy. 


Multiple myeloma is a specific malignant tumor of the bone-marrow 
arising probably from a single cell type, and characterized chiefly by multiple 
foci of origin, a uniform and specific structure composed of plasma-cells or 
their derivatives, rare metastases, albumosuria, and a fatal termination. 
This orthodox definition of the classic picture of the disease is, however, 
subject to extensive modifications as will appear in the context. 

The disease was first described by Mclntyre in 1850 under the term "mol- 
Hties ossium," and has been variously interpreted by subsequent observers, 
as senile osteomalacia (Marchand); myelogenous pseudoleukemia (Zahn); 
primary multiple sarcoma of bone-marrow (Buch). The general clinical 
characters were early recognized, Mclntyre reporting the presence of Bence- 
Jones albumose in the urine, while the tendency of the disease to limif itseff 
to the bone-marrow and the resemblance to an infectious granuloma, have 
been repeatedly noted. Rusticky (1873) seems to have first recognized the 
disease as a specific affection of the bone-marrow and employed the term 
myeloma. Kahler (1889) demonstrated the nearly constant occurrence of 
the albuminuria and the peculiar nature of the protein involved. The 



identity of the tumor-cells in many cases with plasma-cells was pointed out by 
Wright in igoo. The etiology, pathogenesis, origin of the albumosuria, and 
the relation to other tumors and pathological processes in bone-marrow stiU 
remain obscure. 

Gross Anatomy. — The ribs and sternum are the chief original seats of the 
tumors, while the vertebrae, skull, femur, pelvis, and humerus, are less fre- 
quently involved and in the order named. In three cases the ilium appears 
to have been the primary seat (Hoffmann). The tumors form small nodules. 

-Diffuse myeloma of the humerus. 

multiple and a;s large as a bean, or bulky growths as large as an orange, 
associated with many smaller tumors. Twenty-eight separate tumors have 
been observed in the skull (Winkler) , and the diffusion of the process is often 
widespread, involving many bones and revealing a certain systemic quality 
which, however, is nearly always much less marked than that of leukemia. 
Yet the process occasionally appears in diffuse form and circumscribed 


tumors are missing (Winkler, Jochmann, Schumm, Abrickassoff, Hirschfeld, 

The tumors are soft or firm, translucent or opaque, and whitish, gray, or 
deep red according to vascularity. Hemorrhage, infarction, blood cysts, and 
necrotic areas may be present. The bone tissue suffers active absorption 
and the shafts become very thin, or multiple perforations result, or fractures 
occur. After passing the periosteum the surrounding tissues are diffusely 

Many cases terminate without metastases in the organs but distant second- 
ary growths have been found in liver, spleen, kidney, lung and ovary, while in 
certain very malignant cases that probably belong in this category nearly 
every organ in the body may be involved. 

The clinical course of multiple myeloma is that of a progressive malignant 
tumor with gradual and often very extensive invasion of the bone-marrow, 
pronounced lesions in the other organs, a somewhat specific cachexia, and an 
invariably fatal termination. These same general clinical features are 
observed in all types of myeloma, in so-called myelosarcoma, and in endothe- 
lioma of bone, and were closely simulated by a case of chondrosarcoma 
reported by Seegelken. 

The disease arises chiefly in males, in the 4th and 5th decades (24 to 69 
years) and continues a few months or several years. The tumors are usually 
first detected from the pain they produce from erosion of periosteum, pres- 
sure on adjacent nerves, and from general toxemia, and continuous or parox- 
ysmal pain usually persists. Recurring fever has been observed but the 
temperature is usually subnormal. Fractures and extensive deformities of 
the skeleton mark the later stages. Prolonged cases may resemble osteo- 
malacia (Marchand). Anemia may become very severe and exhibit perni- 
cious features (Grawitz), but in many cases the blood shows no pronounced 
changes. Rarely there is an increase of myelocytes in the blood, in one case 
up to 21.8 per cent. (Saltykow, Sternberg). In the blood of Aschoff's case 
Schridde found plasma-cells. Albumosuria has been noted first in the nth 
week or 5th year of the disease. Its occurrence may be continuous or periodic, 
and the quantity of albumose may be scanty or large, up to 6.9 per cent. 
It has been found in less than half the cases, and occurs in myxedema, leu- 
kemia, carcinoma of bone, etc. 

Neither the exact nature, position, nor origin of the urinary protein are 
fully understood. It bears some relation to a severe but non-specific nephri- 
tis which commonly marks the disease, and other proteins may occur in the 
urine. The peculiar qualities of this protein have been fully studied by Mag- 
nus-Levy, Jochmann and Schumm and others. It precipitates at SS°C., the 
precipitate becomes most abundant at 65°, and gradually dissolves at 85° and 
reappears on cooling. Williams finds that its physical properties may change 
from day to day, that it contains a carbohydrate radicle, no phosphorus, and 
falls in the group of mucins. According to Geis the albumose is derived 
from the elastin of bone tissue absorbed by the tumors. 

The spleen is often moderately enlarged, at times partly in compensation 
for the loss of bone-marrow, or from the general toxemia, and from metas- 
tases. The lymph-nodes are rarely hyperplastic (Weber, Scheele, Ewald). 
Extensive amyloid degeneration of the muscles of thorax and intestine was 
■observed by Hueter. The termination of the disease is marked by anemia, 
emaciation, diarrhea, dyspnea, paralysis, and coma. 

Structure. — While in some cases the cells of myeloma exhibit the features 
■of plasma-cells, large or small, with single or multiple nuclei, yet in the entire 
scope of tumors which probably belong in this class the cells vary widely 



in size and character. In some cases the entire tumor is composed of loosely 
packed typical plasma-cells, 8 to 12/i in diameter, round, oval, or polygonal, 
and with opaque, amphophile, non-granular cytoplasm. The single or mul- 
tiple nuclei are relatively small, eccentric, or protruding from the cytoplasm, 
surrounded by a clear zone, and presenting blocks of chromatin arranged 
along the nuclear membrane. Mitotic figures are observed in many cases 
but are less numerous when the cells are typical plasma-cells. Certain cyto- 
plasmic bodies have been interpreted as centrosomes (Christian). The 
clear zone about the nucleus has been described as a vacuolated secretory 
product. In rare cases the round-cells are mingled with spindle-cells which 
probably result from pressure (Wieland). 

In a second .group of cases the cells are larger, giant-cells with multilobed 
nuclei appear, multiple and vesicular nuclei are more prominent, nucleoli are 

Fig. 89. — Structure of a. diffuse lymphocytic myeloma involving whole shaft of humerus. 

(Same tumor as in Fig. 88.) 

large and acidophile, the resemblance to plasma-cells is not striking, the 
tumors are more malignant and metastases occur. Here one finds a structure 
which recalls lymphosarcoma with large cells. 

In this group are rare cases of very malignant large cell sarcoma of the 
bone-marrow with universal involvement of this tissue and very extensive 
growths in lymph-nodes, spleen and other organs, and in which the cells are 
uniformly round, large, granular, and with single or multiple vesicular and 
hyperchromatic nuclei. An extreme case of this group has been fully 
described by Norris. It seems probable that these tumors are a very malig- 
nant form of neoplasm derived from bone-marrow. The relation to myeloma 
is especially indicated by the fact that the tumors in certain regions may 
exhibit a structure closely resembling ordinary myeloma. 

The arrangement of the cells in myeloma is diffuse but there is often a 



scanty stroma separating groups of cells. Occasionally the cells are embed- 
ded in a hyaline acidophihc material, and in malignant forms the tumors 
are very edematous. As the tumor progresses the other elements of the 
marrow disappear, but on the edges tumor and marrow-cells are intermingled. 
Bone trabeculse undergo simple absorption or in the case of compact bone 
osteoblasts are found in lacuna. The tumor never produces bone. An 
inflammatory reaction with exudate of lymphocytes and plasma-cells may 
mark the advance in bone-marrow, periosteum, or other tissues. Some of 
the tumors are oversupplied with small blood-vessels, and in Rusticky's case 
a frontal tumor pulsated. The larger vessels possess well-formed walls but 

Fig. 90. — Structure of plasma-cell type of multiple myeloma. 

many small sinuses occur without definite walls, as in bone-marrow. Hemor- 
rhage is frequent in the central growths and leads to softening, forms blood 
cysts, and leaves pigmentation. Extensive areas of necrosis occur and here 
the cells undergo the usual forms of degeneration. Similar degenerative 
changes overtake single cells in well-nourished tumors and have been used by 
Hoffmann as a basis of classification of cell types. Extensive fibrosis seems 
to be the natural termination of the process in many areas of typical 

"Histogenesis. — The origin of the cells of multiple myeloma has been the 
subject of much discussion which necessarily invades the extensive fields of 
literature on the relations of bone-marrow cells and on the origin of plasma- 
cells (Hirschfeld, Lit.). That the tumor-cells in one group of cases represent 


various forms of plasma-cells is strongly attested by the comparison made by 
Christian. In one group of cases the identity with plasma-cells is so striking 
as to suggest that those writers who doubt this identity have never seen such 
a case. Assuming that the plasma-cells are derived from adventitial cells 
of the vessels the myeloma must be derived from the supporting tissue of the 
blood-vessels of the marrow and not from blood-forming cells. Although the 
earliest stages of formation of myeloma from adventitial cells have not been 
traced, this theory of origin accords with certain important clinical features of 
the disease such as its relatively benign character, and partial resemblance to 
a granuloma and serves to maintain the complete separation of the malady 
from leukemia. It also serves to explain in some degree the capacity of the 
tumor to destroy bone tissue. The origin from adventitial cells nearly 
related to endothelium is further supported by the occurrence of multiple 
perforating tumors of bone cliilically resembling myeloma but composed of 
cells with striking endothelial characters. Such cases have been recorded by 
Kahler, and Marckwald, and observed once by the writer, and in fact if soli- 
tary marrow tumors are included they are not infrequent. 

In another group of cases the cells lack close resemblance to plasma-cells 
and their form and arrangement strongly suggest an origin from the blood- 
fprming cells of the marrow. Both the red- and the white-cell progenitors 
have been included in the scope of this origin. Sternberg, Lubarsch, Herz 
and Charles were able to demonstrate a few neutrophile granules in the cells 
of their cases and this evidence is believed by some to demonstrate an origin 
from the granular leukocyte series (Leukocytoma). Since these granules 
are demonstrated with difficulty in tissues from leukemia and since they may 
well be absent in neoplastic cells, many, after careful study, admit the absence 
of granules, but still claim the identity of the tumor-cells with immature 
myelocytes, basing their conclusion on the other features of the cells (Menne, 

Against this view stands especially the fact that other tumor-like processes 
affecting myelocytes, as in leukemia, pursue a wholly different course from 
myeloma. The so-called myelocytomas being as a rule more active than the 
plasmomas, it may be urged that a greater degree of anaplasia may transform 
olasma-cells into cells resembling myelocytes. 

Opinions regarding the scope of the disease, multiple myeloma, have been 
based not only on the histological structure of the tumors but also upon the 
clinical features. Many observers would adhere narrowly to certain general 
clinical characters which they assume to be essential and would eliminate 
from the group all cases which do not present these features. Thus Winkler 
insists upon the strict limitation of the process to the bone-marrow, the 
absence of metastases, the importance of albuminuria, and the perfectly 
uniform size and character of the cells. Yet it is clear as Lubarsch pointed 
out that the occurrence of metastases is more or less accidental, that albu- 
mosuria may be absent, and the size and form of the cells ma)- vary con- 
siderably, in cases which in general strongly impress themselves as essentially 
the same disease. 

On structure chiefly, Kaufmann separates from multiple myeloma a group 
of malignant lymphosarcomas of bone-marrow. In these the cells are 
larger than plasma-cells, giant-cells occur, a reticular matrix is present, and 
metastases occur. They correspond to the writer's group of more malignant 
myelomas. Yet the cases of Wieland to which Kaufmann refers present the 
main characteristics of malignant myeloma and it is clear that many authors 
have accepted such cases as true myeloma, so that the grounds for their 
separation seem inadequate. Both are multiple primary tumors of bone- 


marrow, perforating the compact tissue, sparing lymph-nodes, producing 
albumosuria, and following the same general clinical course, while varia- 
tions in the cells seem to indicate only different grades of anaplasia. 

A distinctly different condition from myeloma appears to have existed in 
the case described by Hammer in which severe recurrent fever, invasion of 
lymph-nodes, metastatic growths in the dura, and extensive thickening of 
many bones without albumosuria were associated with diiluse lymphoid 
hyperplasia without circumscribed tumors of the bone-marrow. Such a 
picture suggests a true pseudoleukemic process and differs essentially from 
multiple myeloma. Very similar cases of myelogenous pseudoleukemia have 
been described by Baumgarten and by Rubinstein. Hodgkin's granuloma 
affects the bone-marrow but its structure is specific and quite different from 
that of multiple myeloma. 

An erythroblastic origin of certain cases of myeloma is indicated by the 
observation of Ribbert who found hemoglobin in the cells of his case. Norris also encountered such a case. In a recent case of multiple perforating 
bone-tumors with enlarged regional lymph-nodes, I found most of the cells 
smaller than lymphocytes with pyknotic nuclei and strongly acidophile 
homogeneous cytoplasm. 

Finally in certain cases the tumor-cells seemed to the observers to lack the 
.specific features of any of the adult marrow-cells and either to belong to the 
lymphocyte class or to represent an indifferent blood metrocyte. It is nota- 
ble that certain of these cases have been very rapid and malignant forms of 
myeloma (Scheele, Herxheimer, Jellinek, Vignard, Norris). 

Thus if one classifies myeloma according to the views of different observers 
concerning the origin of the tumor-cells the following groups appear: (i) 
Plasmocytoma, (2) erythroblastoma, (3) myelocytoma, adult and embryonal, 
(4) lymphocytoma. Whether such varied interests are actually represented 
in the scope of multiple myeloma or whether we have to deal with varying 
grades of anaplasia in a single cell or origin remains to be determined. At 
present the data seem to favor the former alternative. 

Regarding the etiology of multiple myeloma little can be said. It is a 
disease of middle ages, occurring chiefly in males (19-7), of the middle classes. 
Multiple tumors can hardly be closely connected with trauma, yet three cases 
were referred to injury (Marchand, Ewald, Winkler), and Winkler discusses 
at length its mode of action. Grawitz observed a case which followed 
typhoid fever. Stokvis found two cases in brothers. An infectious origin 
has frequently been suggested from the clinical features, the general signifi- 
cance of plasma-cells, and the relatively slow course of many cases. As a 
late and indirect sequel of an infectious process acting in subjects with 
tissue predisposition the disease is perhaps well conceived. Yet several cases 
have indicated a dependence on nutritional disorders of the bone itself 
which in other subjects leads to osteitis or osteomalacia. 



The term endothelioma is applied to an extensive group of tumors which 
are believed to be derived from the lining cells of blood-vessels, lymph- 
vessels and spaces, subdural spaces, and serous cavities. 

The scope of endothelioma has suffered wide fluctuations because of the 
uncertainty surrounding the embryological relations of different classes of 
endothelial cells, the peculiar position occupied by these cells as intermediate 
between epithelium and fibroblasts, and the difhculty of determining the origin 
of many tumors supposed to belong in this class. Thus some authors find 
the group extremely numerous and complex, while others insisting upon 
more rigid data describe a very restricted class of tumors as endothelioma. 

Origin and Functions of Endothelial Cells. — The embryological relations 
of the flat internal lining cells of the body deserve first consideration in the 
study of endothelioma both for the sake of accurate classification and for 
the bearing on many peculiarities of these tumors. 

Although our knowledge in this field is still incomplete two main facts 
seem to be sufficiently attested: (i) The lining cells of the great body cavities, 
pleura, pericardium, and peritoneum arise from cells lining the coelomic 
cavity which forms early by the sphtting of two layers of mesoderm; (2) 
the lining cells of the blood- and lymph- vessels, as well as the subdural 
membranes are derived from mesenchymal cells and are very closely related 
to connective tissue. 

Through the researches of Huxley, Balfour, Lancaster and the Hertwigs, 
it has been shown that the mesoderm owes its origin in vertebrates to a 
separation of cells derived from both ectoderm and endoderm near the 
primitive groove. The definite layer thus produced splits into two lamellae 
uniting with the ectoderm on the one hand to form the somatopleura, and 
with the endoderm to form the splanchnopleura, and inclosing the coelomic 
cavity between them. In addition to the processes which gives rise to a 
definite middle germ layer, an entirely different process gives origin to the 
sustentative tissues of the body which spread out into the various ramifica- 
tions of the main germ layers. O. and R. Hertwig (1881) and others have 
shown that the sustentative tissues are derived from the middle germ layer 
by the splitting off of many ameboid cells which spread out between the 
epithelial germ layers and form the mesenchyme. From this embryonal 
tissue it is now generally agreed that the connective tissuej and the lining 
cells of the vessels, blood-cells, lymphoid organs, the smooth involuntary 
muscles, heart muscle, and in some cases the striated muscles are derived. 
Concerning the exact mode of origin of the blood and vessels, opinions are 
still at variance but the data are becoming more definite. According to 
Ziegler the fundaments of the blood- and lymph- vessels are derived from the 
interstitial spaces of the primitive body cavity which persists as lacunae 
and interstices in the advancing mesenchyme and become confluent in that 
tissue. A very similar process is described by others who-assume that the 
first vessels arise from mesenchymal cells, which arrange themselves in 



rows and become canalized, forming the vessels. Finally there are those 
who assume an entirely different origin for connective tissues and vascular 
endothelium, attributing the former to migrating cells from the mesoderm 
and the latter to cells of the endoderm which are constricted off as an inde- 
pendent tissue and give origin to the whole vascular system (Uskow, 
Gegenbauer, Ruckert). Thus, embryological data do not warrant funda- 
mental distinctions between the various cells now commonly grouped as 
endothelium since all are of mesodermal origin. Although the coelomic 
lining cells which give origin to. the serous endothehum maintain a lining 
character throughout, that fact does not warrant their identification with 
any forms of epithelium such as is derived from ectoderm or endoderm. 

The vascular endothelium being of mesenchymal origin is closely related 
to the connective tissues and this fact may serve to explain the resemblance 
of certain endotheliomas to connective-tissue tumors, while the difference 
in the embryological history of serous and vascular endothehum seems to 
find some expression in the morphology of tumors derived from these two 
types of cells. Tumors of the serous cells show certain carcinomatous 
rather than sarcomatous tendencies. From the embryologic standpoint 
one might assume a certain relation between endothelioma and tumors of 
lymphoid or smooth rnuscle-tissue, but observation does not indicate that 
this relation is important. In the field of endothelioma as elsewhere the 
adult characters of the cells seem to be of more importance than their embryo- 
logical history. 

The physiology of endothelial cells establishes their position as inter- 
mediate between epithelium and connective tissue. They enjoy certain 
functions recalling those of epithelium but less specialized. In the formation 
of lymph Heidenhain and Hamburger have shown that the lymphatic 
endothehum exerts a secretory action. The secretion of a mucinous lubricant 
is a constant property of serous endothelium. Thoma attributed to a secre- 
tory process the appearance of blood fluids in the closed cell layers of the 
area vasculosa of the chick. Metabolic functions are evidenced by the 
occurrence of fat droplets in the normal endothelium of the pleura (Thoma). 
Phagocytosis by vascular and serous endothehum is often extremely active, 
especially in the spleen and liver. 

Considerable light has been thrown on the origin of endothelium and 
blood-cells by Stockard's study of fundulus embryos without a circulation. 
Here the serous endothelium secretes large amounts of fluid distending the 
pericardium until it deforms the embryo. Vascular endothelium forms 
locally throughout the mesenchyme from wandering mesenchymal cells. 
Similar wandering mesenchymal cells also produce chroma toph ores. The 
vascular endothelium never produces red or white blood-cells. The red 
blood-cells arise from a specialized portion of mesenchyme or intermediate 
cell mass, derived from the posterior region of the embryo. Some of these 
cells wander out on the yolk-sac producing the blood islands of Pander. 
Leukocyte production is largely limited to mesenchymal cells in the anterior 
portion of the embryo but later occurs throughout the mesenchyme, always 
from extravascular mesenchymal cells and not from endothelial cells. 

According to these data vascular endothelium, leukocytes, and red 
blood-cells are separate cell series derived from different mesenchymal 
anlagen. Closely related to them in origin is a series of wandering 

This study is in accord with others which indicate that lymphocytes 
are not derived from the reticulum cells of lymph-nodes. The results are 
probably not irreconcilable with the opinions of Councilman, Mallory and 



others that lymphatic endothelium is a source of some large mononuclear 
cells of the blood. 

The' general pathology of endothelium most clearly illustrates the dual 
tendencies of these cells, revealing on the one hand the assumption of epithe- 
lial ''quaU ties and on the other the possession of certain potencies of con- 
nective-tissue cells. In the course of inflammation of serous surfaces endo- 
thelial cells multiply readily, and form giant-cells and concentric endothelial 
pearls. In granulation tissue they become ameboid and phagocytic. Pial 
tubercles may be composed exclusively of aggregations of proliferating 
endothelial cells. Birch-Hirschfeld found serous endothelium closely related 
to fibroblasts observing that they participated largely in chronic thickenings 

Fig. 91. — Angio-endothelioma arising in a thrombosed vein. 

of the peritoneum. In artificial adhesions of serous membranes, Graser 
and Roloff traced the change of serous cells into fibroblasts while fibroblasts 
in turn produced serous endothelium on new surfaces. Borst saw serous 
cells assume star and spindle shapes, and Lubarsch and Marchand have 
described the formation of intercellular substance and fibrils from these cells. 
Yet Ribbert, Ziegler and others do not accept these observations as proving 
the transformation of serous cells into genuine fibroblasts, and it seems 
highly probable that the new tissues produced by proliferating endothelium , 
have not the permanence of true connective tissue but early undergo 
hyalinosis and atrophy. 

Vascular endothelium also proliferates readily in inflammation and forms 
collections of cells which may resemble the groups of endothelioma or 
carcinoma. In endarteritis the endothelial cells may contribute largely 


to the new tissue, especially in syphilis. In tuberculous endarteritis very 
extensive newgrowth of endothelium is observed in the intima. In the 
organization of thrombi the endothelial cells are very active, wandering out 
into the blood clot, acting as phagocytes, replacing the thrombus with 
strands of spindle or masses of flat cells, and conducting new capillaries. 
The subsequent cicatrization of the organized thrombus illustrates the 
usual fate of new tissues produced by proliferating endothelium. Many 
authors have described the alveolar structures produced by the reversion 
to the cubical type by endothelium of tissue spaces in chronic inflammation. 
In many cases of chronic inflammation of serous membranes groups of 
endothelial cells become incarcerated by adhesions and transformed into 
cubical or cylindrical cells producing structures resembling gland alveoli 
(Ribbert, Renngli, Meyer, Menetrier). In most tissues the endothelial 
cells are the chief source of the epithelioid elements in miliary tubercles. 

In lymph-nodes some remarkable grades of endothelial proliferation 
are observed. In chronic lymphadenitis, probably of tuberculous origin, 
several types of large cell hyperplasia occur, some of which closely resemble 
or actually merge into neoplasms. In scrofulous lymphadenitis the entire 
node may be replaced by masses of poorly staining endothelium. In chronic 
congestion of the spleen the sinuses may appear lined by cubical endothelium. 
These and many other details, which have been extensively discussed by 
Borst, illustrate the behavior of endothelial cells under abnormal conditions 
and they may serve as a standard of comparison in the interpretation of neo- 
plastic processes in these cells. 

History of Endothelioma. — The history of endotheUoma has passed 
through many stages. Arising from a well-characterized cell type these 
tumors are of such wide distribution and exhibit so many of the varied 
qualities and relations of the originating cells that much change and conflict 
of opinion have always existed regarding their scope and classification. 

Koster first clearly maintained the opinion that the endothelium of 
lymph spaces and vessels gave origin to many tumors, which, however, 
he regarded as carcinomas. His demonstration that endothelium may 
give rise to tumors closely resembling carcinoma strengthened the old doctrine 
of Virchow that all cancers arise from the matrix of connective tissue. 
When Thiersch and Waldeyer established the epithelial origin of carcinoma 
and Carmalt showed that cancer-cells could be washed out of the invaded 
lymph-vessels, leaving the endothelial cells intact, the way was opened for 
the recognition of a special group of tumors derived from endothelium. 
Waldeyer's authority encouraged the interpretation of certain tumors in 
various organs as of endothelial origin. 

In 1869 Golgi introduced the term endothelioma, applying it especially 
to psammoma of the dura mater, which has always stood as a standard 
type of the growth. The use of this term became established slowly, many 
authors classifying endothelial tumors with sarcoma. Thus Waldeyer 
described the perivascular growths as angiosarcoma, and Kolaczek in an 
extensive study classified, all the tumors of the endothelium of blood- and 
lymph-vessels as angiosarcoma. The mixed tumors of the salivary glands 
previously regarded as carcinomas were placed by Kauf mann among sarcomas 
until Wartmann in 1879 first maintained that the polyhedral cells of these 
growths are derived from endothelium. Yet the endothelial origin of 
salivary gland tumors has never been universally accepted and while Ribbert 
and many others have firmly maintained their epithelial nature, Kaufmann, 
Borst and others accept in part the endothelial theory. 

As early as 1866 Lucke described a pulsating tumor following a fracture 


of the forearm, the adenoid structure of which he referred to a growth from 
indifferent cells of the connective tissue. Kocher (1868) and Billroth (1869) 
derived somewhat similar tumors of bone from the walls of blood-vessels. 
Tillmanns (1873) and Joppe trace these tumors to the endothelial or peri- 
thelia! cells. Schweninger (1876) clearly traced the development of an 
adenoid bone tumor to the endothelial cells of blood- and lymph-vessels and 
lymph spaces. 

In the orbit Pagenstecher first described a tumor removed by Knapp, 
which by exclusion he was compelled to refer to proliferating endothehum 
of lymph spaces, and this view was endorsed by Knapp (1872) in the descrip- 
tion of later cases. Yet it now appears probable that most- of these tumors 
were derivatives of the lachrymal gland and that endothelioma o'f the orbit 
arises chiefly from the optic nerve sheath. 

In 1879 a group of ovarian endotheliomas was brought forward by 
Marchand, although Leopold had previously regarded certain large solid 
tumors of the ovary as of endothehal origin. In 1870 Wagner recognized 
the endothelial origin of certain pleural tumors and his view was strongly 
supported by Neelsen's study. In 1867 Massey and in 1879 Hubl described 
endothelial tumors of the peritoneal serosa. 

In the skin Kaposi described under the term lymphangioma tuberosum 
a tumor which Jarisch later grouped among the hemangio-endotheliomas, 
but it has long been known that this tumor is derived from the epithelial 
structures of the skin. 

The lymph-nodes were drawn into the sources of endothelioma by Zahn 
in 1874 under the title of sarcoma epithelioides , while in 1881, Hoffmann 
and Schottelius described a similar tumor as endothelioma. Most of these 
tumors appear to have been classed as carcinomas, angiosarcomas and 
alveolar sarcomas. The writer in 19 14 finds that tumors derived from the 
endothelium and reticulum cells of lymph-nodes form separate well-defined 
and rather frequent groups of tumors. 

In the stomach Hansemann maintained that many of the so-called 
scirrhus cancers are derived from endothelial cells of the lymph spaces, 
a view which Borst accepts. Yet this interpretation is open to question 
and the existence of a true endothelioma of the gastro-intestinal tract 
remains unproven if not improbable. Likewise the existence of an 
endothelioma of the cervix uteri is claimed by Amann, Hurdon, 
Kirschgessner, and denied Sperber, R. Meyer, and others. 

In the kidney a group of vascular tumors became generally recognized 
as probable endotheliomas. The common angiosarcoma of this organ was 
interpreted by de Paoli as a derivative of perithelial cells, but the exact 
nature of these tumors and their common origin has not yet been determined. 
Most of them are probably derived from adrenal rests or represent renal 
adenomas. In 1896 Hansemann claimed to have traced the origin of a 
vascular renal tumor to the lining endothelium of a small and originally 
angiomatous growth, but it has remained practically impossible to separate 
from the numerous renal tumors any which may safely be referred to this 

From time to time specific terms have been introduced to designate 
structural types of endothelioma or of tumors so interpreted. In 1856 
Billroth described under the term cylindroma a tumor composed of glassy 
translucent cylinders of hyaline connective tissue between which ran a 
network of tumor-cells. This structure, which is regarded by some authors 
as a form of interfascicular endothelioma, is still often designated as 


V. Ohlen in 1893 pointed out that many authors had been recognizing 
a second form of cylindroma arising by degeneration of groups of cells in 
endothelial masses and the formation of cylinders of mucinous material. 
This structure likewise often received the designation of cyhndroma but it is 
doubtful if it is ever produced by endothelioma. 

For many yfears the term angiosarcoma was loosely applied to certain 
endotheliomas. This term was first employed by Waldeyer and Kolaczek 
to designate all tumors originating from blood- and lymph-vessels and thus 
it came to include those developing from endothelial cells lining the vessels. 
With the recognition that Eberth's perithelial cells, described originally for 
the cerebral vessels, exist also in many other organs, as testicle, adrenal, 
breast, and salivary glands (Arnold, Paltauf, Luschka, v. Brunn), and 
that tumors probably arise from these cells, a specific variety of angiosarcoma 
was admitted under the term "perithelioma." This term was especially 
recommended by Hildebrandt. It has remained firmly established in the 
nomenclature but its relation to endothelioma has not yet been finally 
determined. According to Borrmann, perithelial cells are modified endo- 
thelial cells of adventitial lymph spaces. Many vessels possess not merely 
lymph spaces and perithelium but also a definite plexus of lymph-vessels from 
which Borrmann believes tumors arise. He would therefore distinguish 
between perithelioma and peri-endothelioma, the cells of the former radiating 
from, those of the latter encircling the vessel. In any case perithelioma 
exhibits sarcomatous, not endotheliomatous characters, and in the writer's 
opinion this tumor should be grouped with the sarcomas (Earth, Hippel). 
Thus the old class of angiosarcomas is to-day resolved into perithelioma, 
periendothelioma, and the various forms of endotheliomas arising from 
blood- and lymph-vessels. The present tendency is to limit the term angio- 
sarcoma to cellular angiomas in which the unit is the vessel and not the 
endothelial cell. 

The scope of the group of endothelioma as commonly accepted, especially 
in Germany, reached its acme with the appearance of Volkmann's study in 
which he endeavored to establish the endothelial origin of the mixed tumors 
of the salivary glands. In 1897 Borst reviewed in great detail the behavior 
of endothelial cells under physiological and pathological conditions with a 
result favorable on the whole to the probable wide extent of endothelial 
tumors. In France the tendency to recognize endothelioma as of frequent 
occurrence was much less pronounced (Malassez, Berget, Collet). In recent 
years a rigid critique has emphasized the fact that comparatively few of the 
recognized groups are supported by satisfactory data. In 1899 Hinsberg 
actively contested the endothelial origin of the salivary tumors and he 
has been supported by Ribbert, Marchand, and others, so that the fibro- 
epithelial nature of this group has become generally accepted. Lubarsch 
also has demolished the diagnosis of endothelioma in several types of tumors, 
including Mulert's multiple endothelioma of scalp, a so-called endothelioma 
of the orbit which he derived from the lachrymal gland, and a cylidromatous 
epithelioma of the nasal mucosa. Ribbert does not accept the endotheUal 
-origin of various cutaneous growths, cylindromas, and tumors of parotid. 
The so-called renal endothehoma he thinks is of epithelial origin, while the 
endothelial origin of bone-tumors, as in Sternberg's case, is not proven. 
No tumor of perithelial cells, he believes, has yet been demostrated. 

Thus while the existence of certain distinct classes and many well-defined 
forms of endothelioma was early demonstrated, the attempt to enlarge the 
scope to this group of tumors by assuming an endothehal origin for many 
growths of uncertain nature has met with only partial success. The studies 


of recent years have served chiefly to emphasize the difficulty of separating 
true endothehoma from many typical sarcomas, lymphosarcomas, carcinomas, 
and certain embryonal tumors. 

In this dilemma one may follow Ribbert's plan of discriminating rigidly 
against the tendency to discover endothelial qualities in many tumors 
of uncertain origin. In that case the discussion of endothelioma becomes 
comparatively brief. Or one may accept Borst's conception that the scope 
of endothelioma is probably very wide and one may thus give full expression 
to the views of many who favor the endothelial origin of a very wide variety 
of tumors whose exact nature has not been determined. While fully admit- 
ting the deficiencies of present knowledge my own experience in this field 
inclines me to pursue the latter course. 

Classification and Nomenclature.^ — Various plans of classification of 
endotheliomas have been suggested. Hansemann at one time recommended 
the elimination of the term endothelioma and the designation of these 
tumors as endothelial sarcoma, carcinoma, adenoma, etc. Ackermann 
classified the tumors according to their origin as intravascular, lymphangio- 
matous, and interfascicular (from lymph spaces). Many recognize peri- 
thelioma as a separate variety derived from blood-vessels. 

Borrmann has constructed an ideal histogenetic scheme including 
hemangio-endothelioma, lymphangio-endothelioma, capillary endothelioma, 
perithelioma, and peri-endothelioma. Of capillary endothelioma, with pro- 
liferating capillaries filled with endothelial cells he finds only three recorded 
cases (Langhans, Limacher (Case II), Borrmann). Peri-endothelioma 
refers to a growth of cells of perivascular lymph-vessels concentrically about 
the vessel-wall. In a case described by Perthes he detects a combination 
of perithelioma and peri-endothelioma. However desirable such histogenetic 
distinctions may be, the practical difficulty of unraveling them has stood 
in the way of general adoption of this scheme. Borrmann's capillary endo- 
thelioma is generally, and I think, better designated as an angioma. 

The writer agrees with Borst that the most serviceable grouping of 
endotheliomas is based on their morphology and he has long employed the 
following plan : 

Histological Types of Endothelioma.- — i. Interfascicular.- — The cells 
grow in thin layers between strands of connective tissue. When these 
strands are swollen and hyaline the tumor may be designated as cylindroma. 

2. Alveolar. — The cells grow in small or large groups, as in adenoma. 
When appearing in long sections the groups may be tubular. When mucin- 
ous degeneration occurs in these cell groups a variety of cylindroma is 
produced, but the use of this term for such structures leads to. confusion 
and is inadvisable. 

3. Plexiform. — The cells grow in convoluted columns often surrounding 
vascular paths. Papillary projections may arise from these columns sug- 
gesting the term papillary endothehoma. This structure is represented 
by several tumors of doubtful origin. 

4. Perivascular. — The cells surround definite vascular paths usually 
in concentric fashion, as in the dura mater. Psammoma or sand tumor, 
applies to this growth when the units are calcified. Osteo-endothelioma 
forms rarely in the same manner (Perthes) . 

5. Diffuse. — The cells grow diffusely or without definite arrangement 
in uniform relations. 

6. Miscellaneous changes in the stroma, progressive and regressive, are 
readily indicated by appropriate terms as osteo, myxo, fibro, cysto, etc. 

Perithelioma, whatever its origin, is a highly characteristic structure 


which exhibits the qualities of sarcoma or carcinoma, with which groups 
it should be classed. 

Origin and Growth. — The exact origin of endotheliomas must be referred 
in general to the various endothelial cells, but the earliest stages of growth are 
rarely observed. While the absence of blood in any of the tumor spaces 
distinctly favors a lymphatic origin its presence is of less significance, since 
abnormal communications form between the channels of tumors and the 

According to Borrmann true hemendothelioma is extremely rare and he 
accepts only a few cases, some of which were traced to capillary endothelium. 
The endotheliomas of pleura and pericardium have been referred both to the 
lining cells and to the cells of superficial lymphatics. It seems probable 
that both elements give rise to tumors which differ somewhat in structure. 

The rare and doubtful ovarian endotheliomas have been referred to the 
perivascular lymphatics. While Rosthorn traced capillary blood-vessels 
into the masses of tumor-cells he was uncertain to what extent the lining 
cells of the vessels participated in the process. For the perivascular tumors 
an origin from the lymphatic endothelium lying outside of the lining cells 
must be assumed. Endothelioma of lymph-nodes arises from the lining 
cells of lymph and cavernous sinuses. 

Thus in the great majority of cases it is the endothelium of small vessels 
or lymph spaces which gives origin to endothelioma. Many authors have 
claimed that they could trace tumors to more than one form of endothelial 
structure, as lymph spaces and lymph- vessels, or serous lining cells and endo- 
thelium of lymphatics (Volkmann), lymph spaces and lining cells of blood- 
vessels (v. Ohlen, Franke) . I do not find that the data presented by these 
authors is at all adequate for their conclusions, while some of them were 
probably not dealing with genuine endothelioma. 

From this chief source the growth proceeds either by extension of the 
cells derived from the original territory or by the continuous transformation 
of adjoining normal cells into tumor-cells. Ribbert and Borrmann maintain 
that endotheliomas grow exclusively from their own resources. In various 
endotheliomas Borrmann finds that the proliferating cells of lymph capil- 
laries and spaces distend and enlarge these structures and eventually break 
into other spaces, but they do not produce new capillaries and "contact 
infection" does not occur. On the other hand Best and others describe 
the gradual involvement of normal endothelium in the extension of the tumor 
process. He finds this evidence chiefly along the growing edges of larger 
tumors, while in the very early stages of small growths the increase seems to 
be exclusively in the tumor-cells. The multiplicity of many endotheliomas 
renders the decision in this question unusually difficult. In certain tumors 
of the peritoneum Neelson and Glockner find a progressive involvement 
of neighboring lymphatics, the lining cells of which actively proliferate, 
producing a condition described as lymphangitis carcinomatosa (Schottelius, 
or prolifera (Schweninger). Lubarsch saw this process extend from the 
peritoneum into the liver. It has been held that secondary tumors may 
arise in this way without the transportation of any tumor-cells (Schulz, 
Neelson) . 

In many instances endothelioma appears to invade the tissues chiefly 
or exclusively through preexisting channels, as lymph-vessels and spaces, 
and fails to split up the tissues as do carcinoma and epithelioma. This 
feature was especially notable in an endothelioma of the peritoneum]|^studied 
by Wichern. 

The growth of endotheliomas is usually slow but progressive and in this- 



respect they are comparatively benign. Especially in comparison with 
the forms of carcinoma and sarcoma which they resemble histologically, 
the course of endotheliomas is relatively favorable. When undisturbed by 
the knife the growth capacities of some of these tumors seem to be confined 
within certain limits which do not apply to carcinoma and sarcoma. In the 
dura very slow growth characterizes most of the perivascular tumors and 
pressure symptoms are often long delayed. On account probably of their 
infiltrating character thorough extirpation is often difficult so that endothe- 
liomas are notable for persistent local recurrence. Each recurrence is apt 
to show increasing capacity for growth and changes in structure. This 
fact encourages complete removal of the organ involved and not merely 
extirpation of the tumor. 

Extension to the neighboring lymph-nodes is also late or entirely missing. 
Yet pleural endothelioma has invaded the lung and pulmonary nodes. As 

Fig. 92. — Angio-endothelioma about tendon-sheath. A small, slowly growing tumor. 

an exception to the rule certain endotheliomas of lymph-nodes described 
by Zahn, Volkmann, and the writer were rather malignant. 

Hemangio-endoliielioma. — Many authors have endeavored to emphasize 
and define a distinct group of tumors which they designate as above. (Borr- 
mann, Limacher, Borst). 

Among the tumors commonly included in this group are certain growths of 
the type of angioma but exhibiting an excess of endothelial cells, and certain 
well-known cases are specifically included in the group by the above authors 
(Borrmann, Lit.). Yet Nauwerck's case was designated by the authors as 
a simple hyperplastic telangiectasis. Langhans described a multiple pul- 
sating cavernous angioma of the spleen. Borrmann's case seems to fall in 
the group of simple angiomas with moderate excess of cells, while Steudener 
■describes a rapidly growing ulcerating tumor of the forehead which he 
thought resembled carcinoma. None of these tumors seems to deserve a 
special classification apart from angioma, or perithelioma. 



A second group includes the rare and somewhat obscure intravascular 
endothelioma of corpus cavernosum which cannot be classed with angioma. 
In several small encapsulated slowly growing subcutaneous tumors I have 
been able to trace the abundant cells to proliferating endothelium of dilated 
varicose veins. The circulation appeared to have been previously obstructed. 
The endothelium grows first in broad papillary masses which later become 
fused and enlarged distending the vessel-wall. The endothelial characters 
of the cells were maintained, but the nuclei became enlarged and hyper- 
chromatic. Schlagenhaufer has described such tumors arising in hemor- 
rhoidal veins. 

In the bones, while Narath's pulsating tumor of the tibia belongs with 
cellular angiomas, Marckwald's remarkable multiple tumor of the bones 
may well be designated as an endothelioma of vascular origin. 

Fig. 93. — A type of papillary proliferation of endothelium arising in long distended lym- 

In the testis many authors have described as intravascular endothelioma 
tumors which in all probability are derivatives of teratomas and do not merit 
classification with any form of endothelioma. Of the old angiosarcomas most 
authors are agreed that none can be regarded as hemangio-endothehoma 
and Borst would discard the term angiosarcoma as misleading. 

Thus on analysis it appears that the various tumors which at one time 
or another have led various authors to employ the term hemangio-endothe- 
lioma fall readily into other groups and that this term has a very limited 
application and would include certain rare tumors as follows: (i) The 
doubtful endothelioma of corpus cavernosum; (2) Marckwald's multiple 
endothelioma of bones; (3) a fibrocellular tumor of the thyroid described by 
Limacher in which the vessels were too scanty and the cells too numerous 
for angioma and in which many endothelial cells had become disseminated 


from the vessels (in the ovary somewhat] similar endotheliomas may 
possibly occur) ; (4) a slowly growing tumor of the skin in which a moderate 
number of blood-vessels are lined by many layers of endothelium; (5) an 
intravascular endothelioma arising in hemorrhoidal and other dilated veins. 

Gross Anatomy. — ^The gross appearances of endotheliomas are quite 
as varied as their other characters. Solid circumscribed slowly growing 
tumors occur regularly in the dura. In the serous membranes, the growth 
begins as multiple flat or nodular outgrowths, which coalesce and extend 
gradually over a wide area of the membrane, producing thickenings and 
adhesions. In the pleura the entire surface of both sides may be diffusely 
thickened and the cavities obliterated. In the bones endothelioma tends to 
replace preexisting tissues with little or no increase in the size of the bone 
but with perforations. In the skin endothelioma is said to take the form 
of large single tumors. In the ovary cystic endothelioma is said to occur. 

In the lymph-nodes endothelioma produces solid tumors which may 
reach considerable dimensions. They usually involve more than one node 
and some become systemic. They are often .mistaken for tuberculous 
nodes at first, later for carcinoma. Rarely they are cystic. 

The various forms of secondary changes may chiefly determine the gross 
appearance of the tumor. 

Multiplicity is one of the most constant characters of endothelioma. 
In the dura there may be a score of separate nodules. In the serous mem- 
branes the tumors begin as multiple coalescing nodules. In the spleen 
Weichselbaum observed a notable form of multiple endothelioma or cellular 
angioma. In the bones many have classed multiple myeloma with endothe- 
lioma (Zahn) and a remarkable case of multiple endothelioma involving 
nearly all the bones is recorded by Marckwald. 

Structure of Endothelioma. — The endothelial cell in tumors usually 
retains some of its distinguishing features on which alone the recognition 
of the nature of the growth may often be based. The form is polyhedral, 
often pavement in type, and occasionally cylindrical. Under pressure it 
assumes a spindle form and in edematous tissues it swells to spheroidal 
form and considerable dimensions. 

A well-defined cell-membrane, relatively clear cytoplasm, small pale 
vesicular nucleus with minute multiple nucleoli are features so frequently 
exhibited as to render them valuable diagnostic aids which are too often 
neglected. Especially in separating endothelioma from epithelioma these 
characters are of value since the epithelial cell usually exhibits a granular 
opaque 'cytoplasm, prominent nuclei, and bulky acidophile nucleoli. Yet 
in some malignant tumors the endothelial characters of the cells are lost and 
embryonal epithelium resembles endothelium. In certain endotheliomas 
the cells tend to become flattened and spindle shaped and to form compact 
masses in which hyaline intercellular substance appears in increasing amount 
until finally the tumor is largely transformed into hyaline fibrous tissue. 
This process reveals the natural termination of old endotheliomas of the 
dura. It is sometimes assumed that we have to deal with the production 
of intercellular fibrils by endothelial cells and with the conversion of endothe- 
lial tissue into connective. Yet it appears more probable that the fibrils 
represent elongated and hyaline cell bodies and that the resulting hyaline 
mass represents the gradual conversion of the compressed cells into homo- 
geneous material. Definite formation of intercellular fibrils I have been 
unable to detect in endothelioma. In dural endothelioma their absence 
is conspicuous. 

In various tissues normal endothelial cells show morphological differences 


which dominate the tumors arising in these tissues. Serous endothehum 
is notably flat, bulky, and pavement in t5^e and these features long persist 
in tumors of the peritoneum and pleura. Giant-cells of many forms have 
been encountered in various endotheliomas but their occurrence is compara- 
tively rare (Glockner, Lit.). 

Syncytial masses with scattered nuclei were observed in cyhndroma by 
Sattler. In four pleural' or peritoneal endothehomas Glockner found mono- 
nuclear and polynuclear giant-cells, in one case so numerous as to dominate 
the picture. In an abdominal endothelioma which he derived from the 
lymph-nodes Glockner found very large giant-cells measuring up to 176 
micra. In an endothelioma of the pleura Brosch describes giant-cells 
resembling the Langhans type in tubercle, and I have seen such cells in an 
endothelioma of the scalp. In dural tumors multinuclear giant-cells are 
not infrequent. 

Retained secretion affects the appearance of endothelioma, and endo- 
thelial tumors of lymph-nodes may be cystic. Extensive pigmentation 
occurred in a peculiar tumor with numerous metastases which Rindfleisch 
interpreted as an endothelial myeloma. In an endothelial tumor of the 
sole of the foot attached to the bone I found considerable pigment evidently 
derived from extravasated blood. Secondary changes in the cells and stroma 
greatly alter or dominate the appearance of many of these tumors. Fatty 
degeneration is not prominent but in a case of Ritter's it was quite pro- 
nounced. Glycogenic degeneration has been noted in many cases of actively 
growing endothehoma (Driessen, Volkmann). According to Best its abun- 
dance may distinguish endothelioma as a class from sarcoma. 

Hyaline degeneration frequently affects the cells, stroma, and vessels 
of endothelioma. Involving the stroma of interfascicular growths it pro- 
duces the appearance of cylindroma. In psammoma it is associated with 
calcification of the concentric groups of cells and produces the sand grains. 
Other dural tumors may be rendered quiescent by nearly complete hyalino- 
sis. In a slowly growing angio-endothelioma of dura I found the walls of all 
vessels thick and hyaline and niany occluded by hyaline material, while the 
cells were intact. 

In the true endothelioma mucinous degeneration is of rare occurrence but 
is observed in certain tumors of serous membranes, and in the group of ova- 
rian endothelioma, in both of which it may become extensive with the forma- 
tion of mucinous cysts. 

Calcification overtakes many slowly growing endotheliomas, especially 
the perivascular and psammomatous tumors of the dura. The deposits 
appear in the cells, stroma, and walls of blood-vessels. In a subcutaneous 
tumor of the foot Perthes found extensive calcification leading eventually 
to ossification. In a remarkable form of endothelioma of the orbit I found 
very extensive deposits of calcific nodules producing a form of osteoid tissue. 

The various arrangements of the cells in endotheliomas yield characteristic 
structures which form the basis of the plan of classification previously given. 
These classes by no means exhaust the architectural peculiarities and as in 
carcinoma combinations of several types may occur in the same tumor. 
Likewise epithelioma may be simulated not only by the morphology of 
the cells but by the formation of pearls with hyaline centers. Prickle 
cells and keratohyalin granules, however, are absent in the endothelial 
structures. The structure of sarcoma is simulated by many spindle-cell 
endotheliomas and in this group it becomes desirable to recognize as types 
of sarcoma certain tumors whose cells of origin may be ultimately of endothe- 
lial nature, although modified by location and function. 


Vascular endothelioma tends to remain of smaller size and to assume 
cuboidal or cylindrical rather than pavement types. From the lymph spaces 
develop tumors of rnore embryonal type with cells lacking the above-men- 
tioned adult endothelial characters. These cells are small with opaque 
granular cytoplasm and very intimate connections with the tissue stroma. 
From certain speciahzed derivatives of endothehal cells, as perithelium, 
and the germ center cells of lymph-nodes, the tumors are almost wholly 
lacking in endothelial qualities and constitute subvarieties of sarcoma. 

An intimate relation to the supporting stroma is an important structural 
peculiarity which distinguishes many endotheliomas from epithelial tumors. 
The cells cling to the walls of spaces in spite of the shrinkage of hardening 
and the cell bodies may seem to pass insensibly into the substance of the 
stroma. Small fibroblasts and fine fibrils may ramify between the endothe- 
lial cells. Yet it has repeatedly been shown that these same peculiarities 
may be present in basal cell epithelioma and that they alone do not warrant 
the diagnosis of endothelioma. 

Combination of Endothelioma and Sarcoma. — In some spindle-cell 
sarcomas there is evidence that endothelial cells are commingled with the 
main tumor element but they lose their endothelial characters among the 
derivatives of fibroblasts. Borst mentions the occurrence of large spindle- 
cell sarcomas in which are well-defined groups of neoplastic endothelial 
cells which may resemble carcinomatous alveoh except for the intimate 
relation with the sarcoma-cells. I have seen such appearances in teratomas. 
He also describes angiomatous spindle-cell sarcomas in which the numerous 
vessels are lined by proliferating cubical endothehum recalling gland alveoli. 
In lymphangio-endothelioma the stroma may be very cellular, or one part 
of an angioma may be very vascular and another part chiefly composed of 
cellular stroma. 

Differential Diagnosis of Endothelioma. — Although there have been 
numerous attempts to estabhsh differential diagnostic signs between endothe- 
lioma and other tumors it must be admitted with Lubarsch that reliable 
criteria of general appUcation are lacking. On this account the personal 
impressions of the individual observer have figured too prominently in the 
steady expansion of this group of tumors. 

The most reliable sign of endothelioma is found in the minute characters 
of the cells as above described, but these characters are lost in many actively 
growing tumors and in recurrences. Yet a thorough search for the translu- 
cent polyhedral cells with pale nuclei devoid of definite nucleoli is often 
rewarded in tumors the bulk of which fails to show such cell types. Espe- 
cially in edematous foci the pavement characters reassert themselves. 

The concentric perivascular arrangement is a convincing appearance 
but is limited chiefly to dural growths. All other arrangements of the cells 
exhibited by endothelioma are duplicated by other tumors. 

The occurrence of endothelial pearls with hyaline or calcified centers, but 
without prickle cells and keratohyalin granules, as emphasized by. many, 
stands in the opinion of Lubarsch and Ribbert a strong indication against 
endothelioma. Prickle cells and specific granules are often absent in epithe- 
lioma and with the exception of psammoma I have never seen definite 
pearl formation in any undoubted endothelioma. 

The relation of tumor-cells directly walling circulatory paths is a con- 
vincing feature of some endotheliomas but it rarely exists unless the, charac- 
ters of the cells are even more indicative of endothelioma. The identification 
of lymph paths is usually difficult, while endothehoma of blood-vessels is 
rare. It must be recognized also that the presence of a column of blood 


in immediate contact with tumor-cells is not a certain criterion of endothe- 
lioma. Hemorrhage often fills the alveoli of carcinoma, and in epithelioma 
of the skin blood sinuses invade the epithelial cell groups, lose their endothe- 
lial lining and the blood lies in immediate contact with cuboidal epithehal 

The intimate relation of the cells to the supporting stroma is regarded 
by Borst and many others as a most important criterion. In soft tissues 
the endothelial cells are said to grow out like young capillaries. Yet this 
sign led to error in the case of the spurious endotheliomas of the parotid. 

The desmoplastic properties of many endotheliomas assist in their 
recognition, and a peculiar form of hyalinosis may sharpen this feature, 
as in cylindroma. On the edges of endothehoma the tumor-cells may fade 
in the advancing fibrosis in a manner quite foreign to carcinoma. Fine 
fibroblasts and even capillaries may grow out between the cells of alveoli 
both in endothelioma and carcinoma. The structure of cylindroma, especially 
that form resulting from mucous degeneration in cellular alveoli, is not dis- 
tinctive of endothelioma, since hyaline and mucous degeneration of cell 
groups, stroma and vessels often occurs in carcinomas (Lubarsch, Schmidt). 

On the growing edges the tumor-cells are said to pass insensibly into the 
endothelium of lymph spaces (Volkmann, Hansemann), but this appearance 
occurs also in carcinoma (Ribbert, Borst). Proliferation of endothelium 
of vessels invaded by tumor-cells occurs both in endothelioma and carcinoma 
but rather more in the former. 

The location of the tumor may furnish important evidence, as when 
multiple endothelioma develops in lymph-nodes, bone-marrow, or in any 
tissue not containing epithehum. Yet unless the sources of primary tumors 
are carefully eliminated this evidence is worthless, and there remains the 
possibility of aberrant tissue rests. 

Finally it should be urged that the diagnosis of endothelioma should be 
accepted only when the evidence is clear and conclusive. Otherwise this 
group will continue to represent the resting place of miscellaneous tumors 
on which the data and study have been inadequate. 

Etiology. — In the etiology of endothehoma the influence of chronic 
irritation or trauma and low grades of inflammation must be given a promi- 
nent place. In a few instances it has been possible chnically to estabhsh 
the existence of such factors. Endothehoma of lymph-nodes appears to 
follow forms of chronic lymphadenitis. In the skin chronic trauma is 
abundantly provided, especially in the sole of the foot. In the serous mem- 
branes endothelioma in some cases appears to be connected with tubercu- 
losis. As a rule, however, the tumor appears to develop spontaneously. 
On the other hand it is necessary to assume a local predisposition to tumor 
growth, and in certain cases efforts have been made to establish a relation 
with definite embryogenic disturbances. It cannot be said that much 
success has attended these efforts. Many endothehomas first appear in 
small encapsulated areas suggesting an embryonal isolation of the originating 
tissue. Eberth and Spude have described a small teratoid endothelioma 
of the dura mater. Ribbert suggests that endothehoma of the dura arises 
from groups of cells misplaced during the separation of the original membrane 
into pia and dura. Schmidt has described such collections of superfluous 
endothelial cells in the normal dura. I have observed a small perfectly 
isolated endothelioma embedded in the tissues of the spermatic cord. Yet 
the great majority of endotheliomas fail to give any tangible evidence of a 
relation to embryogenic disturbance. 

The multiple origin of tumors appearing successively over many years 


(Marckwald), the genesis from minute lymph spaces and -vessels, and the 
slow growth, all indicate that these tumors usually develop from adult 
endothelium under the influence of chronic irritation. 


A. Tumor-like Hyperplasia of Endothelium. — Under certain conditions 
endothelium exhibits a grade of hyperplasia which is intermediate between 
inflammatory and neoplastic overgrowth. 

1. In chronic lymphadenitis a peculiar form of endothelial hyperplasia 
may occur which exhibits some of the morphological and clinical features of a 
neoplasm. These cases occur in the group of chronic enlargements of the 
lymph-nodes of uncertain origin and they pursue the course of Hodgkin's 
disease, but specific Hodgkin's structure and specific signs of tuberculosis 
are missing. 

Microscopically one finds irregular foci and broad sheets of large flat 
endothelial cells, obliterating the normal structure, grading insensibly into 
lymphoid tissues, and apparently originating in the sinus endothelium. 
The cells are slightly granular, clearly outlined, polyhedral or elongated, 
with large vesicular hyperchromatic nuclei. The appearance is quite 
different from that of the large cell hyperplasia of tulDerculosis or of the 
common Hodgkin's structure and bears the stamp of a feeble neoplastic 

2. Primary Splenomegaly (T}^e, Gaucher). — This remarkable condition 
consists of an extensive chronic overgrowth of endothelial cells of the spleen, 
lymph-nodes, and bone-marrow, with secondary foci in the portal canals 
of the liver, and is associated with anemia, hemorrhages, and pigmentation 
of skin and other organs. It begins in early life, chiefly in females, commonly 
affects more than one member of a single generation in the family, and pro- 
gresses over many years (39 in Schlagenhaufer's case), proving fatal or 
terminated by complications. The spleen reaches enormous size (6250 
gms. Bovaird) and the liver is greatly enlarged, while the lymph-nodes are 
much less prominent. The pigmentation affects skin, spleen, lymph-nodes, 
and smooth muscle-tissues. 

Microscopically the splenic structure varies in different stages. It 
begins with the appearance of many large clear polyhedral cells lying loosely 
in the pulp tissue occluding and replacing sinuses, cords, and finally the 
Malphigian bodies. Scattered lymphocytes long persist but eventually 
disappear. At the acme of the process, as in Bovaird's case, the tissue is 
composed almost exlusively of closely packed, very large, clear polyhedral 
cells forming diffuse masses or large alveoli. Occasional mitoses may be 
observed and it is evident that the cells are not only well nourished but 
possess considerable powers of growth. Yet the cells lack the hyperchro- 
matic nuclei of tumor-cells and they carefully respect anatomical barriers. 
Necrosis is rare but occurred in Schlagenhaufer's case. In the late stages a 
slow fibrosis overtakes the process and the hyperplastic cells fade into a 
form of hyaline or fibrillated tissue. In the lymph-nodes and marrow much 
the same process is observed beginning in the sinuses. In the liver small 
foci of the large cells appear in the lymphatics of the portal canals. 

From the study of six cases (^Bovaird, Brill, four from New York Hospital) 
I am satisfied that the hyperplastic cells appear first within the pulp cords, 
that they are derived from the lining cells of the splenic sinuses is uncertain, 
that the process originates secondarily in the lymph-nodes and marrow, 
and that the hyperplasia falls below that of a true neoplasm. Whether the 



process in the liver arises in that organ or results from cell emboH appears 

In the etiology a chronic irritant affecting especially the lymphoid organs 
must be assumed to exist. Many features suggest an atypical tuberculosis. 
In two cases (Gaucher, Schlagenhaufer) tuberculous lesions coexisted. In 
Collier's case there was a caseous mass in the mediastinum, while in other 
cases tuberculosis was not eliminated. Lutz found an early stage of the 
process in diabetic lipemia. 

It has recently been shown by Anitschkow that a very similar condition 
of spleen and bone-marrow may be produced experimentally in animals 
by feeding large amounts of cholesterin. The material in the cells does not 
give the staining reactions of myelin bodies, lipoids or fats, nor does it react 

Fig. 94. — Structure of splenic lesion in Gaucher's splenomegaly. Note that the large 
cells lie within the cords and do not line the splenic sinuses. 

to Weigert's myelin stain. Yet it is probably a peculiar lipoid (Schlagen- 
haufer, Lutz). 

B. Endothelioma. — i. Endothelioma of Menin'ges. — This common tumor 
appears in two main forms, (i) As a perivascular endothelioma, and (2) 
as psammoma in which small perivascular units undergo hyahnosis and 
calcification. The tumors are single and as large as a hen's egg, or multiple 
and of miliary dimensions. They first appear as flat elevations on the inner 
side of the dura which slowly enlarge with corresponding displacement or 
atrophy of brain substance. They show little tendency to form adhesions, 
but may be surrounded by a vascular capsule. An infiltration between the 
thickened layers of the dura may occur. They are usually very firm in con- 
sistence, gray in color, with nodular surface or lobulated structure. Strands 
of connective tissue from the dura may be drawn into the growth separating 


lobules of endothelial tissue. The psammomas are extremely hard, usually 
of moderate dimensions and the sand grains are visible in the gross. Occa- 
sionally the tumor becomes pedunculated. On gross inspection the concentric 
units may usually be discerned, but old hyaline forms exhibit a smooth, 
glassy lustre. 

The chief seats are the dura over the convexity, in the falx and tentorium, 
and along the basal vessels, and they may develop over the medulla, and 
down to the cauda. The cerebral tumors have been discussed by many 
authors, while the less known spinal tumors were described in detail by 
Eppinger. Schlesinger collected i8 spinal psammomas located chiefly in 
the central and lower segment. Hippel found a large cellular psammoma 
in the cerebellum and multiple calcified nodules throughout the spinal 
meninges. Small tumors are said to arise from the pia (Borst), and atypical 
forms have been observed in the ventricles arising from ependyma. The 
predisposition of the dura may be associated with the frequent occurrence 

Fig. 95. — Endothelioma of dura mater. 

of "sand grains" on its inner surface producing the condition called by Virchow 
meningitis arenosa. Virchow observed several cases in connection with 
exostoses. Henschen describes multiple endothelioma of the spinal dura 
associated with tuberculous pachymeningitis. 

The structure is quite characteristic and the endothelial features are 
so obvious that this tumor was the first to receive the designation of endothe- 
lioma (Golgi, i86g). The tissue is composed of systems of cells concentri- 
cally arranged about a central lumen, closed or patent, which represents a 
rudimentary vessel. These units may be very small, numerous and sharply 
defined or large and ill defined. Considerable areas may be composed of 
diffuse cells without concentric arrangement, or the masses may be plexi- 
form and convoluted. Borst in one case saw a uniform thickening of the 
meshes of subarachnoidal tissue by cells lying loosely within these strands. 
The cells are polyhedral or of elongated spindle form with the usual endothe- 
lial characters. In some very cellular forms the cell borders are less distinct 
and the nuclei hyperchromatic. 


The origin of the cells has been referred chiefly to the endotheHum of 
perivascular lymph spaces and to perithelium of proliferating vessels, while 
Engert regards them as derivatives of the lining cells of the dura. On the 
other hand M. B. Schmidt has made it appear very probable that most or all 
of the endothelial tumors of the dura are derived from groups of endothelial 
cells which he finds in a large proportion of adults lying in the meshes of the 
pia, near small blood-vessels, and especially in connection with Pacchionian 

In a notable case described by Lindner with multiple tumors of dura and 
similar growths in bladder and urethra, the cell masses inclosed spaces 
communicating with blood-vessels and containing groups of nucleated red 
blood-cells. According to Albrecht's bizarre interpretation of this case 
the tumor was derived from vasoformative cells which retained their capacity 
for red-cell formation. 

The blood-vessels are usually scanty and are carried by the stroma 
derived from the dura. 

Secondary changes are frequent. A form of fibrosis overtakes old tumors 
and may transform them into dense hyaline masses on the edges of which 
islands of growing cells may be preserved. Here the cell bodies become 
fused and hyaline without the intervention of any special intercellular 
substance. Chronic edema may occur in small or large areas in which the 
cells swell to large dimensions and polyhedral form. 

HyaUne degeneration and calcification produce the characteristic struc- 
ture of psammoma. These processes affect chiefly the central cells of the 
concentric units, hyalinosis preceding calcification. Entire cell groups 
may be calcified and almost the entire tumor may be composed of sand 
grains with traces of concentric or irregular striation. The walls of vessels 
may be extensively affected, hyaline changes altering the walls and the 
contents, and calcification appearing in or about the walls and in the con- 
tained material. According to Ernst hyaline material may first appear as 
an intercellular secretion. 

Numerous variations from the typical perivascular structure occur in 
endotheHoma of the meninges. Tumors that have undergone considerable 
fibrosis are often called fibro-endothelioma. The appearance of alveolar 
endothelioma is sometimes produced by cells growing in a fine network of 
connective tissue derived from dura or pia. Very vascular tumors occur in 
which the vessels seem to belong to the tumor process and these may be 
designated as angio-endothelioma. 

In some of the vascular tumors a pecuHar structure is produced by the 
growth of large cells in several layers covering or radiating from the vessels. 
They may involve the pia mater and extend down the sulci. Borst and 
others derive these cells from the perithelium of the blood-vessels and desig- 
nate the tumors as perithelioma. Similar growths arise from the ependyma 
and choroid plexuses where they produce papillary structures. An endothe- 
Homa of the meninges of the dorsal cord, exhibiting the structure of perithe- 
lioma, has been reported from this laboratory by Schlapp. It was rather 
clearly referable to a severe trauma and while it grew rapidly, was completely 
eradicated by curettage. It is difficult to determine the exact position 
of these tumors but in Schlapp 's case the cylindrical and pavement character 
of the cells indicated true endothelial properties in the cells of origin. 

Very cellular growths are frequently described as sarcoma or endothelial 
sarcoma but it was long ago pointed out by Newmann and by Bizzozeroand 
Bozzolo, and lately by Zenoni, that these dural sarcomas are probably derived 
from the endothelial cells. Yet the careful analysis of Lens shows that true 



sarcoma of the meninges arises from the adventitial cells of the blood-vessels 
while an extensive proliferation of arachnoidal endothelium occurs as a 
secondary process which is not neoplastic. The true meningeal sarcomas 
usually take the form of angiosarcoma or large round-cell sarcoma. The 
clinical features of sarcoma are usually missing in the endotheliomas, which 
pursue a slow course without metastases or wide local extensions. 

The course of dural endothelioma is slowly progressive. Pressure symp- 
toms are usually late in appearance and many cases are first discovered 
at post-mortem. Even the vascular growths may fail to alter this course, 
and I have observed an extremely vascular and very cellular diffuse endothe- 
lioma compressing the pons which progressed slowly over a period of. four 


^# -; 



^>\Tf \^ 'fee:- 


Fig. 96. — Slowly growing angio-endothelioma of meninges o£ base of brain. Hyaline and 
calcific material forms the walls of many vessels. 

On the other hand pressure symptoms and recurrence after operation 
marked a case of Taylor's, and a small tumor between falx and convexity 
was found by Lunz in a case of cortical epilepsy. Henschen described a 
remarkable case of multiple endothelioma of the spinal dura associated with 
old tuberculous meningitis, while Dufour observed a somewhat similar 
case in the horse. Metastases are extremely rare but Klebs found nodules 
in the lung in one case, and Lindner is said to have observed a metastasis 
in the bladder. 

Psammoma of nerve-trunks has been described in a considerable group 
of cases; in the optic nerve (Knapp, Ernst, Tailer, Braunschweig, Lit.); 
or acoustic and facial (^Fester) ; but the exact nature of some of these tumors 
is doubtful. 


Acoustic neurofibromas form a characteristic group of intracranial tumors 
and are described under brain tumors. 

Endothelioma of Nerve-trunks and Ganglia. — In many nerve-trunks and 
certain ganglia endotheliomas, usually of the perivascular type, have repeat- 
edly been observed. In some of the common neurofibromas the tumor-cells 
present endothelial characters and are arranged in whorls as in plexifo'rm 
endothelioma. It is difficult to distinguish such tumors from true endothe- 
lioma on the one hand and simple neurofibroma on the other. 

In other tumors of nerve-trunks the structure is that of typical perivascu- 
lar endothelioma as occurring in the dura mater, and psammomatous changes 
may appear in them. 

In the orbit perivascular endothelioma arising from the sheath of the 
optic nerve is a well-defined condition. Of 1 2 extradural tumors of the optic 
nerve collected by Parsons nine were this form of endothelioma. They 
are usually of slow growth, 6 to 8 years, but reach considerable size (Billroth) 
producing exophthalmos. They surround the nerve-trunk in conical form 
or are attached by a pedicle or appear disconnected from the optic nerve, 
in which case another origin may be assumed. The structure shows con- 
centric masses of endothelial cells in which hyahnosis and calcification may 
be prominent. 

Of two orbital tumors belonging to this group and not distinctly connected 
with the nerves, in one I found very extensive calcific deposits among con- 
centric groups of cells showing very hyperchromatic nuclei and evidently 
malignant; in another, 1X2 cm., the cells in broad concentric groups were 
highly pavement in type. Both failed to recur. 

In the extensive literature on tumors of the orbit it is clear that many 
other tumors, especially carcinoma of the lachrymal gland, have been 
interpreted as endothelioma. Thus in a full discussion of orbital tumors 
not connected with the optic nerve Parsons includes many growths of cylin- 
dromatous, peritheliomatous, and sarcomatous character but none of clearly 
endotheliomatous type. Verhoef has shown that the common so-called 
endothelioma of the orbit is a carcinoma of the lachrymal gland. Yet it 
would be unwise to assume that all the suspected endotheliomas of this region 
are spurious. Thus Frank describes a congenital tumor attached to the 
periosteum of the inner wall, containing alveoli of cylindrical cells, diffuse 
cell masses and vessels containing blood and lined by tumor-cells. Even 
here an embryonal carcinoma is difficult to eliminate. In the Gasserian 
ganglion Spiller describes a typical perivascular endothehoma with multiple 
miliary tumors of the adjacent dura. 

Endothelioma of Serous Membranes. — Highly characteristic conditions 
are produced in the pleura and peritoneum by endothehoma of these struc- 
tures and the gross and microscopical features were early recognized as the 
results of a specific process differing from ordinary carcinoma. There are 
such notable differences in the conditions produced in the pleura and in the 
peritoneum as to require their separate consideration. 

(a) Pleural Endothelioma. — The process first appears in the form of 
multiple nodules or flat elevations widely distributed over one or both mem- 
branes. These nodules fuse, the layers of the pleura become thickened and 
adherent and converted into a diffuse firm opaque mass covering and com- 
pressing the entire lung. In Gutmann's case the tumor consisted of a limited 
number of nodules and distinctly papillary masses, and the pleura was only 
shghtly thickened. The costal pleura only may be affected (Scagliosi). 
A sero-fibrinous or bloody exudate commonly accompanies the growth and 
the fluid gathers in small cysts or in larger collections. Invasion of the 


lung may be entirely missing as in Wagner's and Benda's cases, or more often, 
the tumor invades the septa and parenchyma producing irregular nodules and 
masses, and metastases appear in bronchial, axillary, cervical, and medias- 
tinal nodes, and in the liver and spleen. Perls observed metastases in the 
dura and choroid, and Eberth found a second endothehoma primary in the 
pia. The process may extend to the peritoneum or to the pericardium 
(Bassoe), or from the peritoneum to the pleura (Rossier, Pollmann). 

The structure in most cases is rather uniformly that of alveolar and tubu- 
lar endothelioma. The cells are of moderate size, of polyhedral or flat form, 
with hyperchromatic vesicular nuclei and faint nucleoli. They lie in small 
alveoli or long single or multiple rows between cellular or hyaline connective 
tissue to which they are usually intimately adherent. Evidences of secretion 
are usually missing. Borst describes as an endothelioma of pleura a tumor 
containing very large cells closely resembling squamous epithelium and 
forming numerous pearls, but the massing of this tumor about the root of 
the lung and its perforation of the chest wall as well as its structure render 
the diagnosis uncertain. 

The origin of the tumor has usually been referred to the cells of the 
subpleural lymph spaces (Wagner, Bostrom, Neelsen, Volkmann, Adler), 
which have been found affected in very early stages. Malassez, Rossier, 
Pollmann, and others were unable to trace the origin exactly but were sure 
that the tumor did not arise from any epithelial structures. Benda, and 
Guttmann were able to trace the tumor to the lining cells of the pleura, which 
several observers have found to show nodular thickenings due to proliferating 
endothelial plates. All agree that the process begins over a wide area. It is 
possible that two groups of serous endotheliomas should be recognized, one 
invasive with metastases and derived from the endothelium of lymph spaces, 
and another, superficial, nodular, or papillary, and originating from the 
lining cells. 

The clinical course is that of chronic pleurisy with thickening and adhe- 
sions. Rossier especially emphasizes the inflammatory character of the 
process, claiming that it begins as a chronic pleurisy, and the wide extent of 
the growth suggests that it may signify a neoplastic process becoming 
established in predisposed subjects on the basis of inflammatory hyperplasia. 
Perls, Neelson, Glockner, and others even -consider that an infectious agent 
is concerned in the disease and that it is not a true tumor but a form of 
tumor-like lymphangitis. The multiple or diffuse origin and gradual exten- 
sion over normal areas have often been observed and Neelson claims that 
the metastases are not caused by enibolic cells but arise de novo in the affected 
tissue. It may be noted that Brosch in a very early case of pleural endothe- 
lioma observed multiple nodules with caseous centers surrounded by pro- 
liferating vessels and many giant-cells of Langhans type. The resemblance 
to a hemorrhagic tuberculous pleurisy is mentioned by Kaufmann, and the 
difflculty of distinguishing the early nodules from miliary tubercles is noted 
by Wichern. Yet endothelioma of the pleura differs widely from the forms 
of hyperplastic lymphangitis and whatever its etiology the process is neoplas- 
tic. In the diagnosis Fraenkel and Kaufmann mention the importance of 
blood and tumor fragments in the fluid and retraction of the chest wall. 
The disease occurs chiefly in adults but Hibler has observed it in a child of 
five years. 

(6) Peritoneal Endothelioma. — Rokitansky, 1854, described primary 
cancers of the peritoneum composed of multiple translucent nodules lying 
under the serosa and covering many organs. Waldeyer found in primary 
peritoneal growths the structure of plexiform angiosarcoma and cylindroma. 


Birch-Hirschfeld, 1895, distinguished clearly between primary jelly-like 
"angiosarcomas," which he derived from the lymphatic endothelium, and true 
colloid cancer, a bulky tumor which he referred to an aberrant portion of 
intestinal mucosa. A characteristic case of primary endothelioma with 
hundreds of small nodules universally distributed he pictures and describes 
as composed of small groups of endothelial cells which he traced to the 
endothelium of dilated lymph-vessels. 

In 1897 Glockner collected 75 cases of endothelioma of serous membranes, 
pleura, peritoneum, or both, and since that time the condition has been 
generally recognized as the result of a specific process. 

The structure of these tumors generally follows the description given by 
Birch-Hirschfeld and is fully presented in a recent case by Miller and Wynn. 
They found the nodules composed of a framework of connective tissue carry- 
ing numerous vessels and inclosing groups of large polyhedral cells with clear 
cytoplasm. The cell groups often contained mucinous material, and a large 
amount of mucilaginous ascitic fluid was present. Many small clear cells 
and many multinucleated giant-cells were present. The authors trace the 
tumor to the lining cells of the peritoneum, observing all stages of their pro- 
liferation up to the large nodules. They found intimate connections between 
the cells and the accompanying fine connective- tissue fibrils, and describe 
the fibrillation of cell processes leading to the formation of connective tissue. 
Hence they conclude that the tumor has sarcomatous qualities. The intes- 
tinal walls were invaded as far as the submucosa, the general peritoneum 
only in subserous tissue, and the lymph-nodes only in the capsular tissue. 

The ascitic fluid was alkaline, s. g. 1017, when boiled solid still exuded 
a mucinous fluid, reduced Fehling's solution only after boiling with HCl, 
and gave the reactions of albumins which precipitated on half saturation 
with ammonium sulphate. 

A somewhat atypical case is described by Borst. The peritoneum was 
everywhere covered with nodules, plates, and papillary elevations, and numer- 
ous small and large cysts had resulted from inclosure of mucinous material 
not in the tumor-masses but by adhesion of the inflamed peritoneum. The 
diaphragm was perforated and the pleura involved. The cells were flat, 
cubical, cylindrical, or syncytial. The structure was adenoid, interfascicular, 
papillary, and often peritheliomatous. The organs were not invaded but the 
abdominal nodes were extensively infiltrated. The origin seemed to be from 
the subserous lymphatics and Borst suspected that the ovarian region had 
first been involved. 

A still further variation from the usual type is described by Kaufmann 
(I.e. 542) who found multiple nodules, confluent plates, and flat masses 
of partly caseous material surrounded by strata of hyperplastic cells derived 
from the surface endothelium. In many alveoli there were cylindrical 
cells in palisade order. The diaphragm and omentum were chiefly afiected, 
and a hemorrhagic exudate accompanied the process. 

In cases reported by Nager and by Henke the tumor took the form of a 
multitude of miliary or larger cysts covering all the organs, causing universal 
adhesions, and marked thickening of the mesentery and omentum. Some 
of the reports of endothelioma of the peritoneum raise the suspicion that 
the authors are deahng with secondary carcinoma of intestinal origin in 
which the primary tumor has been overlooked. Especially when there is 
marked secretion of mucus the cases closely resemble the so-called pseudo- 
myxoma peritonei which has been definitely traced to the rupture of com- 
paratively benign ovarian cysts. This suspicion attaches to the cases of 
Nager and Hueter, in which the secretion of mucus was abundant, the cells 


cylindrical, and the structure adenocarcinomatous. Very similar conditions 
undoubtedly follow peritoneal dissemination of colloid carcinoma of the 
colon. Further studies in this field are desirable. 

The proper classification of endotheHoma of pleura and peritoneum 
still remains a somewhat fruitless matter of discussion. Ribbert insists 
that the lining cells of the large serous cavities are of coelomic origin and 
epithelial nature and he describes these tumors as carcinoma. Kaufmann 
adheres strictly to principle in separating endothelial tumors derived from 
lymph spaces and carcinoma derived from lining cells, of both pleura and 
peritoneum. The majority of authors, the writer among them, are less 
impressed by embryological considerations than by the morphology and 
physiology of the cells and the general characters of the tumors derived from 
them, and they approve of the classification as endothelioma. The impres- 
sion gained by Birch-Hirschfeld, Waldeyer, and Ziegler that all these tumors 
have sarcomatous qualities is supported by the recent observations of Miller 
arid Wynn of connective- tissue formation by the tumor-cells. I cannot 
trace this process in the few cases at my disposal, but consider that other 
grounds fully support the grouping of these tumors among endotheliomas 
(cf. Monckberg). 

Endothelioma of Bone. — The existence of primary endothelioma of bone 
has been established with difl&culty and not yet to the entire satisfaction 
of some critics. Since this tumor does not always produce structures that 
can be positively identified as endothelial, and since the bone-marrow is a 
frequent seat of metastatic growths, the diagnosis of endothelioma in bone 
requires special caution and the rigid elimination of outlying primary foci. 
This task is often difiicult to accomplish for it is well known that small 
carcinomas of stomach, thyroid, and prostate may yield numerous metas- 
tases in the bone-marrow. Especially do adrenal tumors and carcinomas of 
the kidney show a tendency to form early bone metastases, which may yield 
the first and the chief symptoms of the disease. 

It is further necessary to consider the possibility that primary tumors 
of bone-marrow may arise from embolic cells from normal organs. It has 
long been recognized as probable and Gierke especially has collected much 
evidence to show that cell emboli derived from benign tumors of the thyroid 
and even from simple goiter may lodge in the bone-marrow and give rise 
to tumors. A certain number of tumors of bone described as endothelioma 
are probably of this nature (Gierke, Lit.). A similar mode of origin for 
certain bone endotheliomas which resemble hypernephroma has often been 
suggested. Pick argues that the close contact of the blood-stream with 
normal adrenal cells, and the discovery by Stilling of buds of adrenal cells 
projecting into the veins in the regeneration which follows removal of the 
adrenal in animals, reveal abundant sources for emboli of normal adrenal 

Finally it is possible that misplaced islands of thyroid or adrenal or 
other tissues may exist in the bone-marrow and give rise to tumorS; since 
the wide distribution of accessory thyroid and adrenal tissue is fully 

In view of these considerations the endothehal nature of many old cases 
described as endotheHoma of bone has been disproven or placed in doubt. 

With these precautions it may be stated that two classes of tumors of 
bone deserve consideration as primary endothehoma. 

I. One group of cases consists of single, bulky, circumscribed growths 
which arise from the marrow of long bones, early perforate the bone and 
develop externally. Some become pulsating growths and they may be 



imperfectly inclosed in a bony capsule. Mucoid and hyaline degeneration 
is common and cysts may form. The structure is that of tubular or alveolar 
endothelioma, and often the arrangement of cells follows the type of angio- 
endothelioma. Volkmann describes two typical cases and collects reports of 
others. In none is the endothelial origin proven and later critics have 
specifically referred some of these growths to thyroid or adrenal cells. 

There can be little doubt, however, that there is a primary tumor of the 
bone-marrow of this peculiar struc- 
ture. I have studied six cases of 
this type, three of them located at 
the lower end of the humerus. They 
early perforated the shaft, infiltrated 
the muscles, recurred after opera- 
tion, and produced pulmonary metas- 
tases. The structure presented thin 
strands of connective tissue support- 
ing one or more layers of large clear 
cells inclosing spaces sometimes con- 
taining blood. Or the cell masses 
were compact with occasional central 
necrosis. I have failed to find any 
evidence of bone formation. In one 
case, besides the clear cells there were 
some areas resembling plasma-cells. 

Hansemann, Herxheimer, and 
Schlagenhaufer express the feeling 
that some of these tumors are de- 
rived from osteoblasts, without offer- 
ing definite grounds for this impres- 

2. Multiple small tumors of the 
bone-marrow affecting nearly every 
bone in the body occurred in a re- 
markable case described by Marck- 
wald which stands as a type of 
genuine endothelioma of bone. The 
tumors were most numerous in the 
vertebrae and sternum but many were 
found in skull and pelvis. Some had 
perforated the bone. All other 
organs were free. The patient was 
a male, 53 years, duration 14 months, 
with intermittent fever, pain and ad- 
vancing anemia and cachexia. The 
general condition resembled multiple 
myeloma but the structure of the 
growths was entirely different. The tumors were composed of congeries of 
clear polyhedral cells in diffuse islands and often surrounding capillary 
channels containing intact blood. Neighboring tissue cells were pigmented. 
The diagnosis of intravascular endothelioma seems fully justified. 

Marckwald analyzed several cases from the literature with the conclusion 
that similar tumors have been described by Rustizky, Zahn, Kahler, Wieland 
and Klebs. Some of these, however, were clearly multiple myelomas. 
Howard and Crile describe multiple tumors of the bones which closely 

Fig. 97. — Angio-endothelioma of humerus. 



resemble Marckwald's case and they offer an interpretation of the much 
discussed cases in the Uterature. They conclude that endothelioma arises 
from the endothehum of blood- or of lymph-vessels, and they follow 
Borrmann in assuming that the so-called perithelioma arises from perivascular 
endothelial cells. 

While uncertainty still surrounds the origin of these tumors of the bone- 
marrow it is not clear that they can be separated from multiple myeloma. 
Their multiple character renders an origin from misplaced cell groups of 
any variety highly improbable. 

Periosteal endothelioma with calcification and bone formation has been 
described by Perthes who also finds 30 cases in the literature which he inter- 

FiG. 98. — Structure of angio-endothelioma of bone. 

prets as such. A lymphangio-endothelioma of sternum without bone forma- 
tion is recorded by Coletti. 

Endothelioma of Skin. — In no department has the doctrine of endothe- 
lioma suffered such violent transformation as in the field of tumors of the 
skin. Here several quite distinct neoplasms have gradually acquired a wide 
or general acceptance as endothelioma only to lose their claims for recognition 
in this category through the more careful studies of critical observers. These 
tumors are the lymphangioma tuberosum multiplex of Kaposi, the endothe- 
liomas of the skin of Braun and others, and the soft nevus with its derivatives. 
With the elimination of these groups neither Darier nor Unna admit the 
existence of any genuine endothelioma of the skin. 

Lymphangioma tuberosum multiplex was first described in 1868 by Kaposi 


as a tumor composed of lymph-vessels filled with proliferating endothelial 
cells. Although Jacquet and Darier in 1887 showed that these cell groups 
are derived from the epithelium of dilated sweat-glands, numierous writers 
continued to describe cases as endothelioma. Many authors have pointed 
out their true nature (Frick, Lit.). Stockman has reviewed the history 
of this tumor, fully supported its epithelial origin, and accepted its designa- 
tion as hidrocystoma tuberosum multiplex. 

A second group of tumors of the skin presenting the structure of cylin- 
droma was long regarded as endothelioma. A typical case of this class and 
one much referred to is that of Braun. This tumor was clearly shown by 
Krompecher to be of epithelial origin and is now quite generally assigned 
to the class of epithelioma adenoides cysticum. 

With another group of somewhat similar tumors it has been more difficult 
to reach a conclusion. In 1889 Spiegler described three cases of multiple 
tumors of the scalp which grew over a long period and eventually covered 
the head with numerous discrete tumors resembling small "billiard balls" 
or "tomatoes." The clinical condition was extremely characteristic. Simi- 
lar tumors are also observed in other parts of* the body, especially of face, 
chest, and abdomen. The clinical course, gross section and microscopical 
structure show many resemblances to reticulated epithelioma, with which 
they are classed by Borrmann, Ribbert and others. Mulert's case was 
disposed of by Lubarsch. Many others closely resemble some of the varieties 
of epithelioma extensively illustrated by Coenen. Juliusberger who reviews 
this group at length regards most of the cases as epithelioma, but considers 
Haslund's case and his own as true endothelioma. I do not find satisfactory 
evidence that the tumor described by Juliusberger was derived from endothe- 
lium, bat in Haslund's case the clinical history and structure are rather 
distinct and it is difficult to eliminate an endothehal origin. 

This patient suffered from multiple tumors of the scalp which grew 
rapidly, involved the regional lymph-nodes and proved fatal in 18 months. 
Microscopically the polyhedral or spindle-cells, the formation of long inter- 
cellular fibrils, the absence of definite epithelial characters, the striking 
relation to vascular spaces and channels, and the various stages of tumor 
growth which were traced in great detail from the lymphatic endothelium, 
all accord with an endothehal character. On the other hand their appearance 
as multiple tumors of the scalp of the same general features as the character- 
istic epitheUoma of this region raises the suspicion that Haslund had to deal 
with unusual histological features of an epithelioma. Nevertheless the 
minute histological study of this case brings to light so many endothelial 
characters as to leave in doubt the true nature of the growth, and with it 
the existence of a true endothelioma of the skin. It has already been pointed 
out that the pressure of circulating blood in immediate contact with tumor- 
cells, which is the chief endothehal feature of Haslund's case, is not a rehable 
sign of endothelioma but occurs also in epithelioma of the skin. 

The cellular nevus has been the subject of much discussion and difference 
of opinion. First regarded as an endothehoma derived from lymph-vessels 
(Recklinghausen, Demieville, Lowenbach), the opinion of Unna who inter- 
preted the cells as misplaced embryonal epithelium has found very general 
acceptance by the majority of observers. Yet Kaufmann, Johnston, Schutz 
and others beheve that there are nevi of endothelial as well as those of epithe- 
lial nature. I do not find the distinctions between these groups sufficient to 
warrant any separation of the cases, which seem to me to fall in the class of 
melanoma {q.v!). 

Endothelial psammoma of the skin is described by Winkler. This rare 



tumor occurs in the form of multiple hard nodules in the cutis or subcutis 
which follow the course of nerves and may be traced down to the periosteum. 
They are probably derived from the endothelium of nerve sheaths and traces 
of nerve-fibers may persist. They belong in the group of psammoma of 
nerve-trunks. They consist of concentric masses of endothelial cells which 
are often calcified. According to Winkler certain cases of calcified endothe- 
lioma of the skin (Perthes, Linser, Volkmann) have a similar origin. Psam- 
moma of deeper nerve-trunks is more frequent. 

Angio-endothelioma of the skin and subcutaneous tissue occurred in several 
cases which I have examined. They were slowly growing circumscribed 
tumors of the sole of the foot, finger or face. They were composed of a 

Fig. 99. — Fibro-angio-endothelioma. A circumscribed tumor beneath skin of thumb. 

moderate number of small blood channels traversing wide groups of endothe- 
lial cells. The preponderance of cells over vessels calls for the designation 
of endothelioma (cf. hemangio-endothehoma). 

Endothelioma of Corpus Cavernosum Penis. — This tumor appears in the form 
of multiple nodules or flat infiltrations of the corpora cavernosa behind 
the glans and along the dorsum of the penis. Four cases ,are recorded in 
two of which the course was moderately slow, the growth circumscribed, 
and operation was followed by recovery (Alexander, Hildebrandt) , while in 
the others the disease early involved the lymph-nodes and the patients died 
with general metastases (Maurer, Colmers). The structure consists of the 
cavernous spaces filled with large polyhedral or rounded, rather clear or 
granular cells with large nuclei and prominent nucleoh. The central portions 
of large cell groups are necrotic or occupied by circulating blood (Colmers). 


Invasion of the urethra, septa, skin and especially of the nerve-sheaths is 
observed. Occasionally concentric groups of cells appear. The origin of 
this tumor from endothehum has not been fully determined, but primary 
tumors of the urethra were excluded. Yet Hildebrandt considers the possi- 
bihty, and Borrmann assumes, an origin from urethral epithelium. In 
Alexander's case I must agree with Dunham that the structure was different 
from that of epithehoma and strongly suggestive of endothelioma. 

Creite has described a peculiar urethral carcinoma of undetermined 
origin in a child of two years which may have some bearing on the present 

Endothelioma of Ovary. — Among the tumors of the ovary Leopold (1874) 
described a very large growth which he derived from the endothelium of 
lymph-vessels, while numerous small cysts composing the bulk of the tumor 
he referred to dilated lymphatics. In 1879 Marchand described pap llary, 
cystic, and tubular forms of ovarian tumor which he distinguished from sar- 
coma and considered as very probably derived from endothehum of lymph- 
or blood-vessels. Somewhat similar structures were next observed by 
Flaischlen in a cystic tumor combined with a dermoid cyst. 

A definite relation of the tumor-cells to the endothelium of blood-vessels 
was claimed by Olshausen and Ackermann and by Eckardt, who described 
their tumors as endothehoma intravasculare. On the other hand Pomorski 
and V. Vehts traced cystic and myxomatous tumors to lymphatic endothehum, 
and Rosthorn referred a somewhat different structure, composed of large 
cell groups inclosing blood-masses, to proliferation of perithelial cells. All 
three structures seemed to Amann to deserve recognition as types of endothe- 
lioma. In the disposition of the cells, L. Pick recognized three types com- 
posed of (i) long rows and whorls of cells, (2) cell groups resembling adenoma 
or carcinoma, and (3) diffuse masses of cells resembling sarcoma. In 1896 a 
specific structure, designated as fibrosarcoma ovarii mucocellulare carcino- 
matodes, was described by Krunkenberg and referred in all probability to the 
endothelial cells of this organ (Krunkenberg's tumor). 

Thus such a wide variety of structures of such uncertain origin were now 
included under endothelioma that suspicion arose regarding the validity 
of the rapidly, growing scope of this ovarian tumor. Such obvious combina- 
tions as pyloric cancer and endothelioma of ovary were misinterpreted 
by Rosinski as multiple separate tumors, while Linck even saw in the pyloric 
cancer a metastasis from the ovarian growth. Likewise Krunkenberg 
found the typical structure of perivascular endothelioma in an ovarian 
tumor in a patient who two years later required hysterectomy for a similar 
tumor of the cervix. In Eymer's case also the cervix was involved. 

The first inroad upon endothelioma of the ovary was accomplished by 
Schlagenhaufer, 1902, who showed that the majority of bilateral tumors of 
the ovary are secondary to epithehal neoplasms of stomach, intestine, 
gall-bladder, or uterus, and the typical Krunkenberg tumor he and Glockner 
observed as a metastatic growthi from pyloric cancer with mucus cells. 
Polano (1904) showed that ova-hke structures and perithelioma could appear 
in adenocarcinoma, that metastatic gastric cancer in the ovary produced 
many of the structures of the so-called endothelioma, and that certain mucoid 
endotheliomas were really thyroid struma in this organ. Papaioannou 
also pointed out the close resemblance of metastatic ovarian carcinoma to 
endothelioma and fibrosarcoma. Stone has recently reviewed at length 
the subject of secondary ovarian carcinoma. Glockner denied the validity 
of Pick's histological criteria of endothelioma, especially the absence of an 
intact layer of endothelial cells in the invaded tissue spaces, and claimed that 


the diagnosis of ovarian endothelioma could not be based on histological 
structure alone but required the elimination of other tumors, especially 
of the uterus, by autopsy. Thus grave doubts arose as to the existence of 
any genuine endothelioma of the ovary. R. Meyer was disposed to discard the 
entire group, and Ribbert states (1912) the extreme view that the endothelial 
origin of any ovarian tumor while possible is unproven. 

During this long discussion lasting over thirty years it became apparent 
that unusual difficulties surround the diagnosis of endothelioma in the 
ovary. This organ does not meet one of the conditions laid down by Ribbert 
for the probable endothehal nature of any primary tumor, viz., that the organ 
should not contain epithelial structures. On the contrary, the remarkable 
physiology and changeable morphology of the germ epithelium,. Pfliiger's 
tubes. Graafian follicles, and corpus luteum, offer unusual sources of atypical 
epithelial tumors. The average quota of endothelial cells seems to be defi- 
cient in the ovary if one can accept Polano's observation that no lymph 
spaces lined by endothelial cells exist in this organ. Again it has transpired 
that the ovary is peculiarly liable to implantation metastases by its exposed 
position in the peritoneal c avity, and to deposits by embolism or permeation. 
Finally the doctrine of overgrowth of one element in teratomas has offered 
a peculiar source of certain supposed endotheliomas of embryonal type or 
associated with dermoids (Flaischlen, Vogel) or teratomatous structures 
(Ribbert, Yamagiva). 

One of the most conclusive features of certain endothehomas is the rela- 
tion of cell groups inclosing blood-masses as in angioblastic sarcoma. Yet 
this structure has been clearly traced in the testis to a teratomatous origin 
and an epithelial derivation (Wlassow), and while Sternberg and Monckberg 
have endeavored to establish the existence of a special variety of angio- 
blastic sarcoma or endothelioma apart from teratoma and Pfannenstiel 
recognizes such tumors in the ovary, their teratomatous nature cannot be so 
readily discarded. 

Thus critical research fully justifies the current scepticism regarding 
the endothelial nature of the present group of ovarian tumors and especially 
their origin from adult endothelial cells of any type. 

There is nevertheless a restricted group of ovarian tumors for which 
very strong evidence points to an endothelial origin and their somewhat 
peculiar clinical course, gross appearance, and histological structure warrant 
their separate classification as probable endotheliomas. 

These tumors are usually large, sometimes bilateral (22 per cent. Apelt) 
chiefly solid, yet often cystic. The solid growths show marked lobulation 
and the exterior is nodular, while the general form of the ovary is maintained. 
The cysts are small and numerous as from dilated lymph-vessels, or larger 
and exhibiting on minute inspection papillary outgrowths. A honeycomb 
appearance is noted by Pick. The consistence is firm, and the color varies 
greatly with the blood content. Hydropic areas and hyaline or colloid 
degeneration alter the appearance of the section which in general resembles 

The clinical course is of an actively growing tumor appearing at any 
period of fife but chiefly in adults, v. Velits observed a case continuing 
for nine years, and Kubo twelve years, while Glockner saw a fatal result 
in six months. Local adhesions with ascites and permeation thorough lym- 
phatics are variable but recurrence is frequent. Kubo found no case of 
recovery from bilateral tumors reported as endothelioma but single tumors 
have been successfully eradicated by several operators (Pfannenstiel). It is 
generally agreed that the malignancy less than that of carcinoma. 


The structure falls into several classes (Pick, Pf annenstiel) : 

1. The cells are arranged in orderly rows or whorls between abundant 
fibrils, an appearance which has been likened to strings of pearls. An 
indefinite lumen between such single or multiple rows of small cuboidal cells 
has often been interpreted as a dilated lymph-channel, yet the origin of 
the cells from lymphatic endothelium has never been traced and it is obvious 
that in a large tumor mass original lymph paths are wholly obliterated. On 
the other hand the remarkable long and regular cell columns as pictured by 
Kubo may well signify the reassertion of the physiological tendencies of 
lymphatic endothelium in a desmoplastic tumor. 

2. A definite lumen of small or large dimensions walled by one or more 
rows of tumor-cells and containing lymph or blood characterizes a second 
group of tumors designated as alveolar or intravascular endothelioma. 
The structural evidence of endothelial origin of this group is unusualty 
striking, and there is little doubt that such cases as those of Marchand, 
Pfannenstiel and Kubo are properly regarded as endothelioma. They show, 
moreover, numerous transitions up to structures which really deserve the 
name of angioma. Tumors consisting of one or more large cysts lined 
by layers of endothelial-like cells, described by Schurman and jjy Lange, 
are accepted by Monckberg as genuine endothelioma. If so they represent 
extreme examples of dilated lymph-vessels. Larger plexiform masses of 
cells often form in which small groups of cells surround marrow lumina 
(Kubo) or wide lumina filled with mucinous material appear. Ova-like struc- 
tures formed by mucinous degeneration of cells occur in certain tumors of 
this type whose endothelial nature is doubtful. Similar bodies are found 
in undoubted adenocarcinomas. Hyaline globules in and between the 
cells may also become a prominent feature (Kworostansky). The lymph- 
channels may become dilated and papillary projections of tumor-cells may 
partially fill the small cysts (Eymer), or the channels may be filled with 
blood (Rosthorn). 

3. Diffuse growth of small cuboidal cells forms the bulk of some endothe- 
liomas and appears in portions of many others. Here indications of alveolar 
tendencies may be detected at many points and the tumors bear a resem- 
blance to alveolar sarcoma. The true natuie and origin of such tumors is, I 
think, open to question. 

4. In a group of cases (Kubo, Lit.) the tumor-cells surround vascular 
channels which possess a well-defined wall of normal endothelial cells, giving 
the structure of perithelioma. There seems to be little reason to doubt 
that such tumors, e.g. Amann's case, are of endothelial nature, but the 
attempt to derive the cells from specialized perivascular endothelium has 
not been successful. Pfannenstiel interprets as angioplastic endothelioma 
ovarian tumors described by Michel and by himself in which numerous 
capillary blood channels without walls coursed throughout diffuse cell 
masses. This tumor is probably of teratomatous origin and choriomatous 
type, and while the cells may assume angioplastic qualities the tumor has a 
different significance from the present conception of endothelioma. 

In all of these structural forms the features of the tumor-cells strongly 
indicate their endothelial nature. They are of very uniform small size, 
cuboidal or rounded foim, pale staining cytoplasm, and regular oval nuclei 
with minute nucleoli. 

In the stroma mucinous changes may be prominent or hyaline degenera- 
tion may yield the appearance of cylindroma. 

In the effort to trace the origin the tumors many have assumed that they 
could be subdivided into lymphangio-endothelioma, and hemangio-endothe- 


lioma, intravascular and perivascular, and Kubo has arranged the reported 
cases according to this plan. Its validity, as well as the exact nature of 
many of the tumors included, remain to be estabUshed. Especially for the 
perivascular t}pe the attempt to derive the growth from the hj'pothetical 
perithehal cells deserves little encouragement. Polano could not discover 
any traces of such cells in the ovarian vessels, and it is well kno^Ti that this 
structure may be produced by man)' tumois. 

Endothelioma of Uterus. — In the uterine mucosa and muscularis and 
in the cerAix several tumors have been described as t}-pes of endothelioma, 
and derived b}- the authors from lining cells of blood or hTaph-vessels. 
From the descriptions of these cases I ha^■e not been convinced that a well- 
defined endcthelioma of the uterus has been demonstrated and would agree 
with R. ;Meyer that, while such tumors may occur, it is more probable that 
the alleged cases should be distributed among epitheliomas and sarcomas. 
Clever points out that a considerable grade of endothehal proliferation 
majr occur in the cervix and uterus as a result of inflammation, that the rem- 
nants of Gartner's duct in the cervix have been overlooked as a possible 
source of at}-pical epithelial tumors, and that mam' of the structures of 
endothelioma uteii may be explained as atypical features of other tumors. 

Neither in the clinical course nor in the gross or microscopical structure 
have the reported cases shown uniformity. Thus Pick fimds endothelioma 
in a small portion of the mucosa 'connected with a myosarcoma. Hansen 
describes a large vascular tumor of the m}'ometrium which showed many 
features of an angiosarcoma, and the structure in Silberberg's case falls in 
the same general class. Amann who recorded the first case was content 
to find no connections between the tumor-cells and the surface epithehum, 
but his case and that of Braetz strongly suggest embryonal epithelioma. 
Pohorecky, Kroemer, and Kirschgessner describe structures often duphcated 
by epitheUoma. Hurdon's case was clinically indistinguishable from cervical 
epithelioma with wide extensions, and while portions of the structure suggest 
endothelioma, other areas correspond to epithelicma. Rabb considered 
his case cUnicallj- epitlielioma, and such is the impression clearly conveyed by 
Graefe's report. 

Thus the recorded cases seem to fall into three groups, vascular tumors 
of. the angiosarcomatous type, embryonal epitlielioma, and atj^jical foims 
of ordinary epitheUoma of the cervix. Further critical study would seem 
to be demanded before the existence of endothelioma uteri can be accepted. 

Endothelioma of Stomach. — For many }ears a peculiar condition of cirrho- 
sis of the stomach with p^doric stenosis has been recognized under various 
terms and interpretations. From the surgical side tlie history of the subject 
has been fully reviewed by Lyle who employs Brinton's term linilis plastica 
for tlie general condition. It has long been known that some of these cases 
were atj^pical diffuse carcinoma. In a second group tlie process has been 
recognized as neoplastic and interpreted as endothelioma by Fick, Tilger, 
Jungmann, Cignozzi, Aldegarmann, and others, and Hansemann and Borst 
have lent some support to this view. Szobelew and others have found the 
condition both at the pylorus and in the colon. Yet in none of these cases 
has there been satisfactorv evidence that the process is an endothelioma. 
The structure shows a diffuse and scanty infiltration of large cells with 
hyperchromatic nuclei but of wholly indefinite origin. Alanv of the cells 
ma}' show the mucous degeneration widi signet-ring nuclei of gastric cancer, 
and some small foci of such cells are usually found on careful search. I have 
examined iixe cases of this type in all of which the characteristic mucous 
cells have been found, and in three prolonged search jevealed alveoli of 


typical gastric cancer. Since the typical appearance of endothelioma may 
appear in undoubted gastric cancer there appears to be no ground for the 
assumption that endothelioma of this organ exists. 

That the remaining cases of cirrhosis of the stomach and intestine with 
pyloric stenosis in the adult are the results of a chronic inflammatory process 
has been clearly shown by Krompecher who attributes the changes to 
edema, fibrosis, muscular hypertrophy, and irritation by foreign bodies. 
Recently there has arisen a tendency on the part of clinicians to call many of 
these cases syphilis. 

Endothelioma of Lymph-nodes. — Primary endothelioma of lymph- 
nodes was described by Chambard in 1880 under the term primary carcinoma. 
He recognized a local form of the disease affecting one node or one chain, 
and a generalized form which was always fatal. The structure was alveolar. 
Zahn had previously interpreted an alveolar tumor of the cervical nodes as 
sarcoma alveolare epithelioides. In 1881 Hoffmann and Schottelius used the 
term endothelioma in their report of a case. The endothelial nature of 
these tumors became gradually accepted so that Ziegler, Birch-Hirschfeld 
and Kaufmann admitted their occurrence and referred to single cases of 
their own, and Recklinghausen pointed out certain distinctions between 
primary endothelioma and secondary carcinoma of lymph-nodes. More 
recently quite detailed reports of the gross anatomy, microscopic structure, 
and clinical course of these tumors have been furnished by Ravenna, Gallina, 
Parlavecchio, and the writer. 

In the study of a series of cases collected during a period of several 
years I have drawn the conclusions that endothelioma of lymph-nodes 
is comparatively frequent in occurrence, that the tumors often arise on 
a basis of chronic granulomatous inflammation, that they develop from the 
endothelium of the lymph and cavernous sinuses and are then usually 
classed as secondary carcinoma, that somewhat similar tumors arise from 
the reticulum cells of the lymph-nodes but usually take the form of large- 
cell lymphosarcoma, occasionally of carcinoma, and that the characteristic 
clinical course and microscopical structure of the tumors constitute them 
specific diseases. 

Clinically the disease appears in the form of (i) a systemic involvement 
of many lymph-nodes with a progressive and fatal course lasting from one 
to several years, and (2) as single or multiple tumors of cervical, axillary, 
or other lymphatic chains, often preceded by granulomatous infection, but 
eventually or from the first declaring itself as a locally aggressive recurring 
malignant tumor. 

The systemic form of the disease usually suggests the diagnosis of pseudo- 
leukemia, as emphasized by Parlevecchio. Most of my systemic cases were 
regarded as Hodgkin's granuloma and in some certain of the nodes pre- 
sented the typical Sternberg-Reed histology. In one case of lymphatic 
leukemia, I found alveolar endothelioma of the bronchial nodes. Flournoy's 
case ran a prolonged course of four years and developed cervical, thoracic, 
and very large abdominal tumors. Gallina's patient lived 12 years. 

The localized tumors develop chiefly in neck, axilla, and groin, and are 
commonly regarded as tuberculosis. In one case I followed an originally 
tuberculous infection of cervical nodes recurring repeatedly over a period 
of five years, and gradually assuming the structure of a malignant endothe- 
lioma with elimination of the tuberculous element. Usually the tumors are 
endotheliomatous from the first. In some cases I have been unable to 
separate thel local, cervical or axillary tumors, from a systemic disease of the 
type of Hodgkin's granuloma. The local disease has a lethal tendency, 



but some patients enjoy long immunity after operation, although the ulti- 
mate outcome is usually unfavorable. Zahn and Volkmann describe rapidly 
fatal cases. Extensive tumors of the neck may early cause asphjfxia and in 
others attempted removal proves fatal. 

Gross Anatomy.— The systemic disease is marked by uniform enlargement 
of several -chains of nodes, some of which reach large dimensions and give 
pressure symptoms. The tumors are smooth, at first firm, later very hard, 
usually solid, opaque, and marked by areas of necrosis. Recklinghausen 
saw cystic distension of the abdominal chains, and involvement of many 
retroperitoneal nodes. 

The local tumors are at first circumscribed and the capsules free, but 
later they become large, hard, adherent, and multiple. The infiltration of 
recurrent tumors involves the neighboring structures with compression or 
destruction as in carcinoma, but the progress is usually distinctly slower 

Fig. lOO.^Perivascular endothelioma of lymph-nodes. Nearly all lymph-nodes were 
involved in this case. (Flournoy's.) 

than in epithelial growths. Very extensive fibrosis and large areas of re- 
tained dessicated lymph or necrotic tissue appear in late stages. Almost 
complete sclerosis may overtake some of the tumor-masses. 

Metastases are diflacult to distinguish from systemic extensions, but the 
lesions tend to advance over wide areas, involving thoracic and abdominal 
nodes, and spleen. There appear to be all gradations between the localized 
and the systemic form. Definite metastases were found by Gallina in the 
adrenal, and in the portal canals of the liver by Ravenna. Very extensive 
lymphatic lesions from which skin and muscles were invaded, with edema of 
chest wall, chylous hydrothorax, and metastases in pleura, stomach and 
adrenals occurred in the case of Da Gradi and de Amicis. In Flournoy's 
case a very large tumor, 1780 gm. surrounded the pancreas, nearly all lymph- 
nodes were involved, and there were metastases in adrenal and thyroid. 

Structure. — The structure encountered depends on the stage of the process 



arid on the rapidity of its development. On the one hand it may be difficult 
to separate the early stages from granulomatous inflammation, and on the 
other the late stages resemble carcinoma. The earliest stages of many cases 
appear as a granuloma in which groups or coherent syncytial sheets of sinus 
endothehum or reticulum cells stand out prominently by virtue of their 
size and nuclear hyperchromatism. The cells are large, elongated or polyhe- 
dral with very large vesicular nuclei and diminutive nucleoU, and they 
spring in multiple foci from the lymphatic tissue. The isolation of invading 
tumor-cells is conspicuously absent. The lymphocytes are passive and 
disappear. Ravenna emphasizes the imperfect outlines of the cells and 
their occurrence in broad sheets of cytoplasm interspersed with large vesicu- 
lar nuclei. 


loi. — Diffuse endothelioma of cervical lymph-nodes. 

In some cases fibrosis sets in early and then the tumor-cells become 
isolated and very prominent and the structure resembles alveolar carcinoma. 
Many patent lymph paths lined by tumor-cells may persist or cysts may form. 
Large convoluted masses of coherent endothelium may make up the bulk of 
large tumors in which irregular areas of necrotic tissue containing fatty 
crystals may appear. This structure may be designated as plexiform 

In several cases I have found dilated lymph-channels hned by multiple 
layers of opaque, elongated or polyhedral cells. Or the cells mingled 
with lymphocytes may encircle small arterioles as in perithelioma. These 
structures suggest the term perivascular endothelioma. 

In some cases the whole node is replaced by a diffuse growth of large 
polyhedral or rounded cells which may cohere or fall apart. Giant- 



cells with multilobed nuclei may appear in these tumors, which may be called 
diffuse endothelioma. They may present the appearance of large-cell 

Histogenesis. — Analysis of the origin of endothelioma of lymph-nodes 
involves at once the somewhat complicated questions of the structure of 
lymph-nodes. I have elsewhere discussed this subject in some detail with 
the following conclusions: 

The reticulum cells of the lymph-nodes are derived from connective 
tissue cells, and form a meshwork in which lymphocytes gather from without. 
The reticulum of the follicles and pulp does not produce lymphocytes but these 
two cells constitute separate series (Saxer, Gulland). The reticulum of the 
nodes is not covered by separate endothelial cells, but is itself a cellular 



Fig. 102. — Diffuse endothelioma of lymph-nodes. Coherent masses of large cells with 

clear cytoplasm. 

syncytial structure which incloses spaces continuous with the lymph paths 
of the node. In the cavernous and lymph sinuses the lining cells have 
pavement qualities of adult endothelium, while in the cords and follicles the 
reticulum cells retain more of their original mesoblastic tendencies. Hence 
tumors aiising from sinus endothelium may be expected to show pavement 
or epithelial characters, while tumors of reticulum cells should exhibit 
some of the mesoblastic features of their parent cells. 

To a large extent these expectations are realized in the study of the 
tumors. To the sinus endothelium has been traced the origin of several 
tumors presenting alveolar or perivascular structures composed of polyhedral 
cells (Ravenna, Ewing). The reticulum cells of the lymph cords are 
found multiplying in the early stages of many tumors, especially in the 
outlying nodes, and it is probable that changes in the form of these cells 



toward adult endothelium accounts for the preponderance of large polyhe- 
dral cells in the whole group of tumors. From the reticulum cells of the 
follicles 1 have repeatedly been able to trace the development of large round- 
cell tumors of the type of large round and giant-cell lymphosarcoma, which 
tumors must be separated from lymphocytoma. 

That the tumors are not secondary carcinomas is sufficiently attested 
by the failure to find any primary tumor and by the very wide diffusion 
and slow course of the systemic disease. That they do not arise fiom 
aberrant epithelial rests within lymph-nodes has been proven by the detec- 
tion of early stages throughout the whole node or several nodes of a chain 

Fig. 103. — Perivascular endothelioma of cervical lymph-nodes. 

With this complex morphology and histogenesis it is difficult to establish 
a legitimate nomenclature. The terms employed, endothelioma of various 
types, emphasize the form of the tumor-cells and their probable origin. 
The tumors of loose round-cells derived from the reticulum of the follicles 
differ markedly from the others and it seems unwise to urge any change 
in their usual designation as large-cell lymphosarcoma or reticulum-cell 

Etiology. — Many observers have been impressed by the apparent depend- 
ence of endothelioma of lymph-nodes upon lymphatic infection. The clini- 
cal features of the systemic cases closely resemble those of pseudoleukemia 
or of Hodgkin's granuloma. Parlavecchio stated that the histological 
structure revealed the transformation of an inflammatory into a neoplastic 




Fig. 104. — Unit structure in a tumor of lymph-nodes. Large cells growing in and about 

small blood-vessels. 



. ( 

9 '-' 

*•'. i»* !i.\ 


Fig, 105. — Early stage of diffuse endothelioma of lymph-nodes. These cells are derived 
from the reticulum and sinus endothelium. 


hyperplasia of endothelium and he clearly pictured these transitions. In 
the recurrences of one case I was able to trace an original tuberculous granu- 
loma into the neoplasm with elimination of the tuberculous structure. I 
have reported several other cases in which various phases of remarkable 
endothehal overgrowth . occurred in tuberculous or other granulomatous 
lesions. Many of the localized tumors present in outlying nodes the changes 
of frank or atypical tuberculosis. Most of the localized tumors and the 
advanced systemic lesions fail to show any traces of granuloma, suggesting 
that the tumor soon overgrows the granuloma or arises immediately in 
response to the infection. Other infectious agencies besides tubercle prob- 
ably figure in the disease, and of these Hodgkin's granuloma requires first 
consideration. Syphilis has not thus far been implicated. 


That a definite group of tumors exists which deserves recognition under 
the term perithelioma must be admitted, but the exact origin of these tumors 
and their relation to other neoplasms are questions requiring more extensive 
data than we now possess. 

The existence of a special type of cell about the blood-vessels of the pia 
mater was first shown by Eberth. In salt solution he was able to isolate 
a thin membrane composed of polygonal or multipolar cells surrounding 
these vessels. These cells were distinct from the true adventitial cells and 
formed an outer boundary between the vessel-wall and the perivascular 
lymphatics. Although Eberth did not clearly separate these perithelial 
cells from the endothelium of perivascular lymphatics, and KoUiker (1896) 
describes only perivascular lymphatics about cerebral vessels, similar cells 
have been described about the blood-vessels of the hypophysis (.Waldeyer), 
carotid gland (PaJtauf), adrenal, breast, salivary glands (Eberth, Arnold, 
Sertoli, Luschka), and testis (Henle). Nevertheless the existence of such 
cells has not been accepted as a well-established fact in normal histology, and 
the derivation of an important group of tumors from them rests on an un- 
certain basis. The belief that certain tumors do arise from Eberth's peri- 
thelia! cells is suggested chiefly by the peculiar arrangement of the tumor- 
cells, radiating out from the supporting vessels, but it must be admitted again 
that all the tumors in this class do not exhibit this peculiar arrangement 
of cells. Indeed one must agree with Ribbert that no tumor has yet been 
clearly traced to perithelial cells. 

Notwithstanding these obscurities the group of perithelioma maintains 
its existence for the following reasons, (i) These tumors present as a structural 
unit the blood-vessel, and it is clear that they originate from some cells in 
the walls of the vessels. (2) Although originating from blood-vessels they 
differ in structure from angioma on the one hand and from endothelioma on 
the other. (3) Their gross and clinical characters differ from those of an- 
gioma and endothelioma. 

From this point of view the term perithelioma cannot be regarded as 
fully applicable, and many other terms have been employed by different 
authors, as angiosarcoma (Kolaczek), plexiform angiosarcoma (Waldeyer), 
perivascular endothelioma, and peri-endothehoma (Borrmann). Yet many 
types of sarcoma are derived from blood-vessels and it seems better to 
limit the term angiosarcoma to the cellular and malignant forms of angioma. 
The term endothelioma is inadmissible since the perithelial tumors differ 
notably from recognized types of endothelioma and they have not been 
traced to an endothelial origin. 


The position of perithelioma seems to be intermediate between sarcoma, 
especially angiosarcoma, and endothelioma, and atypical and diffuse forms 
tend to vary in either direction. Thus diffuse perithelioma may be difficult to 
distinguish fiom large spindle-cell sarcoma, while certain cases which it is 
difficult to eliminate from the peritheliomas (Paltauf, Hildebrand) resemble 
alveolar endothelioma. 

An important source of confusion which has been prominent in the history 
of angiosarcoma and perithelioma arises from the fact that many other 
actively growing and malignant tumors assume the structure of perithelioma, 
either in the original tumor or in metastases or recurrences. Hence Roussy 
recognizes perithelioma only as a peculiar morphological structure without 
histogenetic significance. In primary carcinoma of liver and in the round- 
cell teratoid carcinoma of testis a peritheliomatous structure is frequently 
observed. In the third recurrence of a teratoma testis I found the entire 

hA '^'■ 

^^se^^'V ^-^^S^ ip-^^C 

'.— ^ 

Fig. io6. — ^Low magnification of cross-section of perithelioma. 

tumor composed of blood-vessels sheathed with large radiating cells. For 
such structures the term secondary perithelioma may be suggested. Again 
the tissue giving origin to a malignant epithelial tumor may be very vascular 
so that the tumor from the first assumes the form of spurious perithelioma. 
These conditions exist in kidney and adrenal, epithelial tumors of which 
have often been described as peritheUoma. Indeed it is open to question 
if a considerable number of supposed primary peritheliomas may not owe 
their structure to the excessive development of blood-vessels supporting 
rapidly growing embryonal epithehal tumors. Metastatic melanoma has a 
persistent tendency in this direction, so that one may well look for active 
pigmented moles in all cases of perithelioma of groin or axilla. 

Gross Appearance.— Perithelioma usually appears as a well-circumscribed, 
very vascular, at first slowly, later rapidly growing tumor which may reach 
considerable dimensions. It is very subject to hemorrhage, necrosis and 



ulceration. In early stages it may be freely movable in the tissues exciting 
little local reaction, and may be firm, soft or almost fluctuating or pulsating. 
On cross-section the tissue appears to be composed of thick-walled and 
variously twisted blood-vessels. This appearance strongly indicates that 
the predominant unit in the tumor is the blood-vessel and since this unit is less 
pronounced in secondary and spurious perithelioma its presence is the best 
indication of the nature of the growth. Local extensions and metastases 
occur late but after imperfect removal, recurrence, rapid extension, metas- 
tases and cachexia supervene, so that the general prognosis is unfavorable. 
Structure. — (i) The typical perithelioma presents a very characteristic 
structure. It is composed of a congeries of small vessels of the type of 
arterioles sheathed with large polyhedral fusiform or rounded cells which 

Fig. 107. — Topography of a true perithelioma. 

radiate in numerous layers from the vessel-wall. At a certain distance from 
the vessels nutrition often fails and the cells undergo anemic necrosis which 
serves to isolate and emphasize the vascular units. In many tumors the 
vessels disappear, the growth is of diffuse large spindle-cells and the tumor 
resembles simple sarcoma. Giant-cells of small and large dimensions with 
multilobed nuclei are a frequent and characteristic feature, especially in the 
diffuse growths. Intercellular substance appears in many of these tumors 
and completes the resemblance to sarcoma. 

2. An alveolar arrangement of distinctly polyhedral or cylindrical cells 
appears in another group of peritheliomas, especially in tumors of the carotid 
gland, bone, and skin (Paltauf, Hildebrand, Hanke). Since the cells remain 
polyhedral or rounded, and intercellular substance is scanty or absent, these 
tumors resemble endothelioma. At present their exact position must remain 

In both structural varieties of perithelioma, the walls of vessels and the 


stroma and tumor-cells may undergo hyaline and especially mucoid degenera- 
tion. Some tumors described as myxosarcoma are probably myxomatous 
forms of perithelioma. The hyaline degeneration may produce the structure 
of cylindroma. 

Clinical Types of Perithelioma. — i. The subcutaneous tissue is one of 
the chief seats of typical perithelioma. The loose tissues of the anal region, 
orbit, axilla, popliteal space, and regions of embryonal fissures, but not the 
more exposed regions, furnish the exact points of origin, and this distribution 
suggests a relation to embryonal disturbances of superfluous tissue. 

The deep fasciae and intermuscular septa are sometimes the point of 
attachment of peritheliomas which soon appear to be subcutaneous, as 
described by Kolaczek, Tillmanns and others. 

The buccal mucosa over the hard and soft palate is the seat of a well- 
defined clinical group of typical perithelioma which has been emphasized 
by Eisenmenger (Lit.). 

2. The pia mater is the seat of a group of somewhat atypical perithe- 
liomas. A well-circumscribed growth 5 cm. in diameter, displacing the brain 
tissue, with cells concentrically arranged about numerous blood-vessels 
as in angioma, was early described by Arnold. A flat tumor of the dura 
mater of the base, involving roots of several cranial nerves and composed of 
many small vessels lying among loose round and polyhedral cells, was inter- 
preted as a peri-endothelioma by Wells. It contained calcific areas and bone 
spicules. In the brain substance perithelioma has been described by Wald- 
eyer and Billroth, and many others, and Waldeyer showed that these tumors 
had previously been mistaken for primary carcinoma of the brain (Eberth, 
Arndt). Probably the majority of bulky sarcomas of the brain substance 
present this structure. 

3. In the bones perithelioma appears as a soft vascular tumor, often 
pulsating, which early perforates the shaft, may invade the surrounding 
tissues and tends to produce distant metastases (Jaffe). Hildebrand has 
collected a series of cases and describes one in which the congeries of vessels 
was plainly evident in the gross. The origin he finds to be in the bone 
substance and the growth is of ten preceded by trauma or fracture (Kocher, 
Lucke). Howard and Crile interpret nine cases in the literature and three 
of their own as arising from perivascular endothelium which they identify 
with perithelioma. The cells are epithelioid in form and their arrangement 
is either alveolar or tubular. It is not certain that all the tumors described 
under this term have been clearly separated from endothelioma or from 
epithelial tumors derived from aberrant cells groups. Many have been 
described as angiosarcoma, and before the studies of Billroth and Waldeyer 
they were commonly regarded as primary carcinoma of bone. They are 
probably identical with angio-endothelioma. 

4. In the breast perithelioma is one of the common forms of vascular 
sarcoma. It occurs chiefly in adults after 50 years as a single circumscribed 
small or large tumor movable in the breast but usually adherent to the skin. 
The density, retraction of skin, and involvement of lymph-nodes of carcinoma 
are absent. Growth is slow but recurrence after operation is frequent 

In the ovary, among the many forms of so-called endotheUoma there is 
one described by Amann which has certain features of perithelioma. 

5. In the kidney, PaoH and others have described large vascular tumors 
which they interpreted as perithelioma, but it seems highly probable that 
these growths should be divided among adenocarcinoma of the kidney and 
hypernephroma. The cells exhibit epithehal characters. The cases of 


angiosarcoma of the liver described by Arnold exhibited the structure of 
this type but whether they are primary or secondary remains to be shown. 
In cases of primary carcinomas of the liver, areas of typical perithelioma are 
very frequent (L'Esperance). 

6. Perithelioma of Carotid Gland. — Since the appearance of the simul- 
taneous studies of Marchand and Paltauf describing a characteristic tumor 
of the carotid gland, many other cases have been reported, minute analysis 
of .the structure of these tumors has repeatedly been made, and the normal 
histology and embryology of the gland itself has been thoroughly investigated. 

While Marchand suggested the origin from perithelial cells, and Paltauf 
fully accepted this view, later observers have been less confident regarding 

Fig. io8. — Structure of a carotid gland tumor. 

the exact nature of the growth. Kauffmann employed the term alveolar 
tumor, and compared it to a hamartoma, in which all elements of the gland 
participated. The complex structure of the organ, as shown by Kahn, 
Stilling, and Monckberg, may serve to explain many peculiarities of the 
tumors and the difficulty of devising a satisfactory designation. While the 
tumor belongs in a class by itself, the term perithelioma recalls a prominent 
structural feature and has been very generally employed. 

These tumors arise usually about puberty and slowly reach sufficient 
dimensions to protrude from the side of the neck, producing a deformity which 
has been pictured in detail by Keen and Funke. Their location is first 
upon and behind the bifurcation of the carotid artery which is often inclosed, 
compressed and even invaded by the growth. Prolongations run in various 



directions, upward toward the skull, inward to phaiynx (Reclus), and down- 
ward. Most of the tumors are rounded, lobulated and encapsulated. Rup- 
ture of the capsule may occur (Monckberg, Leithoff). They are firm or 
soft, may yield a bruit, frequently pulsate, and Reclus was able to largely 
obliterate a tumor by compression. On section they are vascular, grayish 
red, or slightly brownish from chromafiine substance in the specific cells. The 
markings show vessels, plexiform cell masses, concentric structures, and 
hyaline areas. 

The growth is slow, occupying years. Heinleth observed a very large 
tumor existing over 30 years. While numerous adhesions form and vessels 
may be infiltrated and penetrated, a skilful enucleation is usually successful 
in preventing recurrence, but Licini estimates the mortality from the opera- 
tion at 35 per cent. In several cases recurrence has followed operation, 

Fig. 109. — Structure of a carotid gland tumor. 

and the cervical lymph-nodes have been found involved (Kaufmann, Kretsch- 
mar, Kopf stein). General metastases are not observed. 

The structure is in general that of an alveolar perithelioma but varies 
in many details. Post-mortem changes occur rapidly and only prompt 
fixation of operative material in Zenker's or Orth's fluids preserves the 
true structural detail. 

The chief cells are distinctly polyhedral, granular, and may contain chro- 
mafi&ne substance. They are arranged in compact groups without lumen 
but surrounded by hyperplastic endothelium lining the numerous capillaries. 
Much larger cell groups are often observed, the central portions of which are 
prone to degenerate and form cavities in which blood may be found. Definite 
capillaries may also traverse the cell groups (Monckberg). Hyaline degen- 
eration of the cells gives rise to intracellular globules, and affecting the stroma 


produces peculiar knobs projecting into the capillaries (Kaufmann). The 
numerous vessels may produce a cavernous tissue or they may be much re- 
duced in hard tumors. They are lined by prominent endothelial cells which 
sometimes grow out into the vessels in syncytial buds, or mingle with the 
chief cells of the tumor, producing complex pictures. The stroma may be 
scanty and acellular or hyaline, or infiltrated with tumor-cells, and Leithoff 
describes a case in which he regarded the stroma as sarcomatous. 

These and many other structural details which appear in many combina- 
tions in the reports of numerous authors show that the tumors of the carotid 
gland are sui generis. Both the specific cells and the endothelium paitici- 
pate in the tumor process, yielding organoid growths involving the entire 
gland. Although Monckberg thought he could identify ganglion-cells in one 
case, the numerous non-medullated nerve-fibers of the normal gland appear 
only in traces and do not participate in the tumor process. Since the tumors 
arise about the period when a physiological regression usually overtakes the 
gland, some disturbance in the course of this regression may be an etiological 
factor (Heinleth). Advanced osteomalacia occurred in a case of Oberndor- 
fer's, and various authors mention other possible etiological factors. 



Two important physiological properties influence the conception of 
tumors of lymphoid tissues. 

First, lymphoid tissue responds to irritation with inflammatory hyper- 
plasia far more actively than any other tissue. Second, lymphoid tissues 
are relatively mobile rather than fixed, and lymphocytes are not only ame- 
boid but are structurally placed in easy access to lymph and blood paths. 
Hence tumors and tumor-hke processes in lymphoid tissues are frequent and 
as a rule tend to become widely diffused. These same factors render unusually 
difficult the distinction between simple and neoplastic hyperplasia of lym- 
phoid tissue, by introducing special standards, both clinical and anatomical, 
in the interpretation of lymphatic hyperplasia. 

For these reasons the discussion of lymphoma in general necessarily 
includes the consideration of various processes, some of which are chiefly 
inflammatory, some neoplastic, and others intermediate in position. Lack of 
knowledge of etiological factors, the intricate relation of many forms of 
lymphoid hyperplasia, and the occasional transformation of one process to 
another, render this method of treatment practically advantageous and 
theoretically sound. 

Extensive tumor-like hyperplasia of lymphoid tissues occurs in the 
following scheme of clinical conditions: (i) Simple lymphoma; (2) leuke- 
mia, lymphatic and myelocytic, in its various phases; (3) pseudoleukemia; 
(4) Hodgkin's disease; (5) lymphosarcoma. 

The complexities of the subject of lymphoid tumors depend chiefly upon 
the lack of knowledge of etiology and partly upon the lack of accurate ana- 
tomical classification. It seems to the author that some advantage would 
result from the rigid application of simple anatomical principles in the classi- 
fication of these processes, even at the risk of carrying the anatomical dis- 
tinctions too far. 

Anatomically there are three elements in lymphoid tissues capable of 
originating lymphoid tumors, viz.: (i) lymphocytes; (2) reticulum cells of 
follicles and pulp, and (3) endotheUal cells of pulp and cavernous sinuses. 
Connective tissue cells, endothelium of vessels, and cells of vessel-walls do 
not appear to be concerned in the common tumors of lymph-nodes. 

Classified according to present indications of histogenesis, tumors of 
lymphoid tissue appear as follows: 


Reticulum cells 

Endothelial cells 

Anatomical Type 

Large round-cell hyperplasia 
or neoplasia 

Endothelial hyperplasia or 


Clinical Type 
Simple lymphoma 
Tuberculous lymphoma 
Lymphatic leukemia 
Malignant lymphocytoma 

Granuloma malignum 
Myeloid leukemia 
Hodgkin's sarcoma 
Large-cell lymphosarcoma 

Endothelial hyperplasia 

of tuberculosis, etc. 



Several difficulties stand against the finality of this classification. 

I. The relation between lymphocytes and reticulum cells is not fully 
determined. While the older anatomists, headed by Flemming, held that 
reticulum cells of the follicles become transformed into lymphocytes, prac- 
tically all the recent work indicates that the two types of cells are entirely 
separate (GuUand). 

It would greatly simphfy the questions of histogenesis if separation of the 
two types of cells were estabUshed. Most pathologists, however, the writer 
among them, are unable to satisfy themselves that reticulum cells during the 
course of inflammatory and neoplastic changes may not assume the form of 
lymphocytes. Yet the separate nature of the two cell types is of such great 
importance in the study of lymphoid tumors, that it seems wise to accept 


no. — Multiplication of follicles in simple chronic lymphadenitis, in postpyloric 
nodes draining a chronic ulcer of stomach. 

the dictum of embryology until it is proven incorrect. In most pathological 
processes the two cell types undoubtedly maintain their individuality. 

2. The distinctions between reticulum cells and sinus endothelium are 
much less definite. In fact, in the development of lymph-nodes both arise 
from modified mesoblastic cells. Yet in a study of endothelioma of lymph- 
nodes the writer has brought evidence to show that the reticulum cells 
give origin to the large-cell lymphosarcomas, while the sinus endothelium 
yields tumors of a different structure. 

3. While the lymph-nodes in lymphatic leukemia contain an excess chiefly 
of lymphocytes, the origin of the larger round-cells in the nodes of myelocytic 


leukemia is not satisfactorily determined. Their classification as derivatives 
of reticulum cells must be regarded as tentative. 

4. If a complete solution of the etiology of lymphoid tumors were provided 
it might destroy much of the value of an anatomical classification. The 
best known excitant of lymphoid enlargements, tuberculosis, produces over- 
growth of lymphocytes, extensive or exclusive hyperplasia of reticulum 
cells and endothelium, and also introduces many blood-cells by exudation. 
Hodgkin's granuloma exhibits similar or identical tendencies. 

Yet there is much value in classifying lymphatic tuberculosis from the 
anatomical standpoint, and similar value may remain for the anatomical 
subdivisions of other diseases when their etiology is known. For the definite 
tumors of lymph-nodes this objection loses much or all of its force, but it is 
difficult to determine when one is dealing with a granuloma or a tumor of 
this tissue. 


Local or regional, circumscribed, chronic enlargements of lymph-nodes 
which are self limiting and do not lead to systemic invasion are rather 
common pathological conditions (Le Count, Lit.). They appear in the 
neck, axilla, groin, subcutaneous tissue, submucous tissues, parotid and 
other glands, and occasionally in other regions. They form solid masses 
of tumor-like dimensions and firm consistence, and on section exhibit a 
smooth opaque uniform surface often presenting firm, light, more prominent 
areas, and revealing little or no tendency toward degeneration or necrosis. 
They may be composed of a single node or a congeries of nodes with fused 
capsules. After reaching a certain limit of giowth they may remain station- 
ary for months or years producing only local symptoms. The tendency 
toward regression is slight. I have observed a simple lymphoma of four 
years duration occurring beneath the skin of the scapular region which failed 
to show signs of regression or fibrosis. The prolonged course is a feature of 
many cases, that of Le Count continuing for 15 years but still showing 
hyperplastic germ follicles. The tumor described by Wagner and Hewitt 
reached dimensions of 7 X S cm. in six years. Internal growths of uncertain 
history may be found at autopsy. 

The structure varies at different periods but is dominated by diffuse 
overgrowth of typical small lymphocytes. In early stages lymph follicles 
are increased in size and number, are readily visible in the gross and exhibit 
enlarged germ centers, the pulp tissue is composed of closely packed lympho- 
cytes, the sinuses are indistinct or obliterated, the vessels compressed or 
packed with lymphocytes; or the normal landmarks are obliterated by a 
diffuse growth of lymphocytes and the distended capsule is infiltrated. 
. Simple lymphoma may be separated from leukemia by the blood test, and 
from pseudoleukemia by its local character, but from neither on the histolog- 
ical structure alone. From lymphosarcoma the diagnosis may be difficult, 
and should be based on the absence of the usual signs of malignancy, as rapid 
growth, mitotic figures, active destruction of trabeculae and capsule, and 
atypical quality of the cells. 

However obvious may be the inflammatory origin of many cases of 
simple lymphoma, the excessive hyperplasia, prolonged course, and often the 
apparently idiopathic origin renders the condition of interest in the field 
of tumors and requires the recognition especially in chronic cases of certain 
features of true neoplasms. 



I have observed one case, structurally requiring classification as lymph- 
adenoma, persisting for iS years and eventually proving fatal with rapid 
local growth and asphyxia (Figs, iii, 112, 113). 

In the etiology of simple l}'mphoma many forms of chronic bacterial 
infection aj)pear to be concerned, especially tuberculosis. According to 
Paltauf the regeneration of l_\-mphocytes following an acute infection may 
be excessive and ])rolonged, giving rise to a tumor-like overgrowth which 
persists long after the removal of bacteria and their products. In the case 
of tuberculosis, it has been shown by Bartels and others that the structure 
of simple h-mphonia may be exhibited by h'mph-nodes harboring viable 
but attenuated tubercle Ijacilli, and the frequency of simple l)-mphoma with- 
out tubercle bacilli in tuberculous subjects suggests that this lesion may 
be one of many which seem to be referable to atypical tuberculosis. Such 


Fig. 1 1 1 . — .Malignant lymphaJenoma. 

cases may be designated as paratujjcrculous lymphadcjiitis (Adami). In 
certain cases it is impossible to establish any relation whatever to tuberculo- 
sis, or anv acute infection, or any chronic disturbance in neighboring mucous 
membranes or tissues. 


The leukemic jjrocess in its entire sco])e illustrates many phases of hy])er- 
plasia of lym])hoid tissue. This disease may exhibit as its anatomical 
basis: (a) A transient leukemic blood picture with slight hyperplasia of the 
cells of l\-mph-nodes or marrow; (b) acute leukemia associated with acute 
hyperplasia of l\-mphoid tissues; (c) chronic leukemia, with leukemic lympho- 



mas and hyperplasia of bone-marrow; {d) chronic leukemia with partially 
neoplastic hyperplasia of lymphoid and myeloid tissues; (e) chloroma. 

While it is impossible to prove that leukemia is always caused by a single 
infectious agent or that it represents anything more than a'clinical entity 
comparable to leukocytosis, these various lesions belonging to essentially 
the same condition illustrate every gradation from an acute inflammatory 
process up to a chronic and somewhat maUgnant neoplasm. 

The lesions of leukemia both lymphocytic and myelocytic, involve the 
bone-marrow, lymph-nodes, spleen, and the other preexisting lymphoid 
tissues, while heterotopic growths of lymphoid tissue appear in the liver, 


Fig. 112. — Malignant lymphadenoma. Same as Fig. iii. 

kidney, lung, skin, serous membranes, and many other organs. The picture 
is that of a diffuse systemic involvement of blood-forming organs with 
secondary invasion of other tissues. 

In the bone-marrow the process begins in multiple hyperplastic foci 
which enlarge, coalesce, and eventually extend the limits of lymphoid marrow 
throughout all portions of the skeleton. ,At the height of the process the 
new tissue is firm, light colored, opaque, or pyoid, or beset with foci of 
infarction, necrosis, hemorrhage or mucoid softening. Fat cells, sinuses, 
and all other normal structures of the marrow are commonly obliterated by a 
diffuse growth of hyperplastic round-cells. Banti describes an initial hypere- 
mia with widening of sinuses and atrophy of fat-cells. Islands of hyper- 



plastic marrow-cells then appear in and about the sinuses and these are 
foUowerl lj_\' pure lymphoid growths starting from tine walls of arterioles and 
gradually inwading the entire marrow. The spongy trabecuUe are often 
absorljed and even the shafts ma\' be thinned, fiul a (listinctl\" aggressive 
destruction of bone as in true tumors is missing. In rare cases of myelemia 
there is wides])reacl osteosclerosis (Schmorl). 

In the lymph-nodes the disease appears in one or more chains spreading 
to other chains and involving all the iiTrles of the chain simultaneously. 
Bulk}- tumors are thus produced in which all the nodes remain discrete 
or the capsules are fused by inflammatory reaction. The process in the 
lymph-nodes differs in the two main types of leukemia. In myelemia the 
pulp cords lirst show increase in size b_\- the aj^pearancc of m>'elocytes which 
gradually distend the cords, obliterate the sinuses, cause passive atrophy 

Fig. m. — Jlalignant Ij.'mphadenoma of cervical nodes. Detail of Fig. in. 

of the lymph-follicles, and e\-en infiltrate the capsule or the surrounding 
fat tissue. Giant-cells and nucleated red blood-cells may be present, but 
in the more pronounced lesions onl\- large myelocytes are present. 

In h-mphemia the condition in the lymph-nodes \-aries. In some cases 
there is h>-pertrophv of lymph follicles which become greath- enlarged and 
fusing with the pulp cords convert the entire node into diffuse h'mphoid 
tissue. In such nodes the lesion is primary and these cases may be desig- 
nated' as lymphoplastic lymphemia (Pappenheim, i\Ie}-er, Heinicke, 
Hirschfeldj. In other cases the lymph-folUcles fail to enlarge but seem 
passively compressed b\' new growth of lymphocytes beginning in the pulp 
cords Opinions differ' as to the relati\-e frequency of the two types of 
lesions According to Walz, Pappenheim, Grawitz, Kelly, Gulland, and 
Goodall, Ivmphemia is always primarily myelogenic, while Ehrlich and Pm- 



cus maintain that the whole lymphoid system may be primarily involved. 
My own cases seem to illustrate both conditions but I have not seen active 
growth of germ centers in lymph-nodes of leukemia. The extent of the 
affection of lymph-nodes varies greatly. Thus Walz found scanty changes 
in nodes, much in spleen and marrow; Pappenheim describes absence of 
changes in nodes, few in spleen, much in marrow: while Meyer and Heinicke 
record early and extensive lesions in nodes, later in spleen, and later and only 
focal hyperplasia in bone-marrow. Turk ventures to conclude from clinical 
signs that marrow lesions may be absent in early stages of lymphemia and 
this view is substantiated by Rosenfeld and others who believe that the 
marrow lesions may appear late in the disease. 

The splenic lesions are very similar to those of the lymph-nodes. The 
organ exhibits a diffuse enlargement, a smooth opaque surface and section, 
areas of infarction and necrosis, but seldom any locaHzed tumor growth. In 
myelemia there may be a simple myeloid transformation of pulp with atrophy 
of foUicles. Here myelocytes, giant-cells, nucleated red blood-cells and 

Fig. 114. — Advanced chronic myelocytic leukemia. 

phagocytes are mingled, much as in the marrow. More often there is a 
diffuse infiltration with cells of the myelocyte class. In late stages and 
especially in very large spleens the cells are greatly reduced and fibrosis, 
hyahne degeneration, hemorrhage and necrosis are prominent. In lymphe- 
mia the foUicles may be intact and enlarged but more often they are oblit- 
erated by a diffuse growth of lymphocytes which lie loosely in the disordered 
cords and sinuses. 

In the submucous lymphoid tissues the process resembles that of the 

The heteroplastic deposits of leukemia occur chiefly in the liver. Here 
there is a diffuse growth of round-cells in the capillaries which has little 
tendency to destroy liver cords but miliary or larger lymphomas occur in 
the portal canals. A certain grade of these lesions is practically constant 
in all forms of leukemia. In lymphemia the portal nodules are more exten- 
sive than in myelemia. 

In the kidney leukemic infiltrations, focal or diffuse, frequently occur. 
The cutaneous manifestations of leukemia are varied and complex (Oertel, 



Lit.). Lymphemia has been observed with spindle-cell sarcoma of the skin, 
with multiple lymphomas, and with mycosis fungoides (Nekam, Pincus, 
Lit.). The characteristic lymphomas appear as multiple nodules chiefly on 
the face and neck or widely distributed as a papular eruption. They may 
be preceded by or associated with other eruptions. The older lesions may 
ulcerate (Nicolau). They are composed of lymphocj^tes grouped about 
the vessels of the corium. 

Cutaneous lesions occur with moderate frequency but in great variety, 
and illustrate in their course many phases of a granulomatous inflammation 
unfolding into a tumor-like process. The lesions include petechije and 
hematomas, vesicles and bulls, macules, papules, and pustules, and localized 



-Bone-marrow in lymphatic leukemia. Focal hyperplasia of small lymphocytes 
in center of femur. 

or difl'use tumors. There may be marked pruritus and terminal bronzing 
(Hazen, Lit.). Very similar conditions occur without leukocytosis, as in 
Hodgki'n's disease, pseudoleukemia, and mycosis fungoides. In rare cases 
the lesion is diffuse and universal (Arndt). Other organs and tissues are 
occasionally the seat of leukemic infiltrations. 

There has been much discussion regarding the origin and nature of the 
heteroplastic deposits. Perhaps the majority of observers have concluded 
that in both forms of leukemia the hepatic deposits arise from hyperplasia 
of preexisting lymphoid structures in this organ which undergo simple 
hyperplasia in lymphemia and myeloid changes in myelemia (Sternberg, 


Some authors assume the existence of latent myeloid foci m liver, lung. 



etc., which serve as a source of the new tissue in myelemia (Dominici, 
Hirschfeld, Walz). From these foci may arise the myeloid changes so often 

Fig. ii6. — Lymph-node in chronic lymphatic leukemia. Note the very uniform size and 

distribution of the cells. 

Fig. 117. — Chronic myelocytic leukemia. Section of lymph-node showing diffuse growth 

of cells of uniform size. 

seen in certain infectious diseases attended willr anemia (Meyer, Heineke). 
Others accept the local origin of the heteroplastic lymphomas of lymphemia, 



but deny the existence of myeloid foci. Thus Sternberg distinguishes 
sharply between lymphemia and myelemia in the origin of their heteroplastic 
deposits, and Meyer and Heineke deny that any myeloid foci exist in normal 

Fig. 1 1 8. — Structure of spleen in a case of lymphatic leukemia. Focal and diffuse 
overgrowth of lymphocytes, which are found in cavernous sinuses, pulp sinuses, pulp cords, 
and in enlarged follicles. 

nodes, spleen, or liver, although they find the source of lymphemic growths 
in slumbering lymphoid nodules. At the other extreme Banti holds that all 
heteroplastic leukemic nodules represent metastatic growths of embolic 


^ ^ 

x^ y- .«.< ^'^ 

Fig. iig. — Liver in lymphatic leukemia. Lymphocytic infiltration of portal canals and 

hepatic capillaries. 

At present it seems impossible to offer proof of the entire validity of any 
of these theories. I believe, however, that too much importance may be 



placed upon the hypothetical lymphoid foci of such organs as the liver and 
kidney. In the normal adult human organs such foci are conspicuously 
wanting, but in lymphemia such organs rapidly develop lymphomas. For 
the more distant lymphomas of skin, brain, and serous membranes, espe- 
cially those which reach a bulky growth, it is more difficult to accept an origin 
from preexisting lymphoid foci, while the origin from embolic leukocytes seems 
much more probable. 

Finally in those cases in which the lesions exhibit more definite neo- 
plastic properties, as in the leukosarcomatosis of Sternberg and in chloroma, 
it is highly probable that the laws of neoplasms control the development of 
the heteroplastic nodules. Hence it is necessary to assume that in mild 

Fig. 120. — Miliary lymphocytomas in the liver of lymphatic leukemia. 

forms of lymphemia the heteroplastic nodules arise in loco while in the 
more intense grades of leukemic hyperplasia they are chiefly embolic in 
origin or else one must accept their embolic origin throughout. The latter 
position seems to me the more logical and more in accordance with observed 

In myelemia the evidence in favor of the embolic origin of heteroplastic 
foci is much more direct. Myeloid foci do not exist in normal adult liver, 
or spleen, and while they develop there in infectious diseases and in anemia 
their constituent cells have not been satisfactorily traced to preexisting 
cells in these organs. In leukemia the blood-vessels in and about myeloid 
areas often contain thrombi of myelocytes and giant-cells which offer the 
readiest explanation of heteroplastic nodules (Meyer, Heineke, Schwartz, 



Michaelis). Yet an embolic origin does not involve a neoplastic nature, if 
as Sternberg claims, leukemic cells are normal marrow-cells produced in 
large numbers, for sarcoma-cells are atypical and grow without normal 
restraint. Yet Ribbert and others find that the leukemic cells are undiffer- 
entiated and abnormal and exhibit in some degree the characters of tumor- 
cells. In ordinary myelemia the myeloid areas show little of these neo- 
plastic properties, but in the atypical and rare forms bulky tumors arise, 
tissues are invaded, cells are atypical, and neoplastic characters are rather 

Fig. 121. — Leukemic deposits in the heart-muscle. Similar tumors in spleen and kidneys. 
Enlargement of thymus and many lymph-nodes. {After Seelig.) 

Changes in Bleed. — Closely dependent upon the organic lesions are the 
changes in the blood, the full discussion of which falls outside the scope of 
this work. Briefly they consist in extensive increase of leukocytes and pro- 
gressive anemia. In lymphemia the cells are small and medium sized lym- 
phocytes in chronic cases, and larger undifferentiated lymphocytes or mononu- 
clear cells in acute cases. In myelemia the new cells are chiefly neutrophile 
myelocytes, many of which are of abnormally large size, atypical eosinophile 
myelocytes, excess of polynuclear neutrophile and eosinophile cells, mast 
cells, and various degenerative forms. While in typical cases the leukocytosis 
is the first demonstrable symptom, it varies greatly in degree, in its relation 
to the extent of the organic lesions, and in its permanence. While the 
leukocytes may outnumber the red cells, leukemia occurs with smafl numbers 
of atypical cells in the blood, and their numbers are subject to extensive 
variations from numerous factors. 


The cause of the access of large numbers of cells to the blood-stream s